BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191

[email protected] https://web.facebook.com/OMEGACENTER2014

𝓐– 𝟏 • 𝑫é𝒇𝒊𝒏𝒊𝒕𝒐𝒏 ∶
𝑺𝒐𝒊𝒕 (𝛀, 𝓕, 𝓟) 𝒖𝒏 𝒆𝒔𝒑𝒂𝒄𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒔é. 𝑶𝒏 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒔𝒖𝒓 (𝛀, 𝓕, 𝓟)
𝛀⟶ℝ
𝒕𝒐𝒖𝒕𝒆 𝒂𝒑𝒑𝒍𝒊𝒄𝒂𝒕𝒊𝒐𝒏 𝑿 ∶ , 𝒗é𝒓𝒊𝒇𝒊𝒂𝒏𝒕 𝒍𝒆𝒔 𝒅𝒆𝒖𝒙 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏𝒔 𝒔𝒖𝒊𝒗𝒂𝒏𝒕𝒔 ∶
𝝎 ⟼ 𝑿(𝝎)
𝓪 • 𝑳′ 𝒆𝒏𝒔𝒆𝒎𝒃𝒍𝒆 𝒅𝒆𝒔 𝒊𝒎𝒂𝒈𝒆𝒔 𝑿(𝛀) ⊆ ℝ 𝒆𝒔𝒕 𝒖𝒏 𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆 𝒅𝒆 ℝ , (𝒐𝒏 𝒑𝒆𝒖𝒕 𝒂𝒗𝒐𝒊𝒓
̅ = [−∞, +∞] = ℝ ∪ {±∞})
𝑿(𝛀) = ℝ
𝓫 • ∀𝒙 ∈ ℝ , 𝑷𝑿 ({𝒙}) = 𝟎
𝑹𝒆𝒎𝒂𝒓𝒒𝒖𝒆 ∶ 𝑷𝑿 ({𝒙}) = 𝟎 , 𝒑𝒂𝒓 𝒄𝒐𝒏𝒕𝒓𝒆 𝑷𝑿 (𝒙 ∈ [𝒙, 𝒙 + 𝒅𝒙[) ≠ 𝟎

𝓐– 𝟐 • 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶

̅.
𝓪 • 𝑫é𝒇𝒊𝒏𝒊𝒕𝒊𝒐𝒏 𝟏 ∶ 𝑺𝒐𝒊𝒕 𝑿 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒔𝒖𝒓 𝑿(𝛀) ⊆ ℝ
𝑼𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒇𝑿 ∶ ℝ ⟶ ℝ 𝒆𝒔𝒕 𝒂𝒑𝒑𝒆𝒍é𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒔𝒊 𝒆𝒍𝒍𝒆 𝒆𝒔𝒕 𝒑𝒐𝒔𝒊𝒕𝒊𝒗𝒆 𝒆𝒏 𝒕𝒐𝒖𝒕
+∞
𝒑𝒐𝒊𝒏𝒕 𝒕 ∈ ℝ 𝒐ù 𝒆𝒍𝒍𝒆 𝒆𝒔𝒕 𝒅é𝒇𝒊𝒏𝒊𝒆 , (𝒇𝑿 (𝒕) ≥ 𝟎), 𝒊𝒏𝒕é𝒈𝒓𝒂𝒃𝒍𝒆 𝒔𝒖𝒓 ℝ 𝒆𝒕 𝒔𝒊 ∫ 𝒇𝑿 (𝒕)𝒅𝒕 = 𝟏
−∞

𝓫 • 𝑫é𝒇𝒊𝒏𝒊𝒕𝒊𝒐𝒏 𝟐 ∶ 𝑳𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒓é𝒆𝒍𝒍𝒆 𝑿 𝒔𝒖𝒊𝒕 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒇𝑿 𝒔𝒊 ∶

𝒃
∀𝒂, 𝒃 ∈ ℝ ; ∀𝒃 > 𝒂 , 𝑷𝑿 (𝒙 ∈ ]𝒂, 𝒃]) = ∫ 𝒇𝑿 (𝒕)𝒅𝒕
𝒂

𝑹𝒆𝒎𝒂𝒓𝒒𝒖𝒆 ∶ 𝑫𝒆𝒖𝒙 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒑𝒆𝒖𝒗𝒆𝒏𝒕 𝒂𝒗𝒐𝒊𝒓 𝒎ê𝒎𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é

𝒔𝒂𝒏𝒔 ê𝒕𝒓𝒆 é𝒈𝒂𝒍𝒆𝒔.

𝓐– 𝟑 • 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 ∶
𝑶𝒏 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝑿 , 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝑭𝑿 𝒅é𝒇𝒊𝒏𝒊𝒆
𝒙
𝒔𝒖𝒓 ℝ 𝒑𝒂𝒓 ∶ ∀ 𝒙 ∈ ℝ , 𝑭𝑿 (𝒙) = 𝑷𝑿 (]−∞, 𝒙]) = 𝑷𝑿 (𝑿 ≤ 𝒙) = ∫ 𝒇𝑿 (𝒕)𝒅𝒕
−∞

 𝑳𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅′ 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝑿 𝒆𝒔𝒕 𝒄𝒓𝒐𝒊𝒔𝒔𝒂𝒏𝒕𝒆

𝒔𝒖𝒓 ℝ , 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆, 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆𝒎𝒆𝒏𝒕 𝒅é𝒓𝒊𝒗𝒂𝒃𝒍𝒆 𝒔𝒂𝒖𝒇 𝒑𝒆𝒖𝒕 ê𝒕𝒓𝒆 𝒆𝒏 𝒖𝒏 𝒏𝒐𝒎𝒃𝒓𝒆 𝒇𝒊𝒏𝒊 𝒅𝒆 𝒑𝒐𝒊𝒏𝒕𝒔
𝒆𝒕 𝒐𝒏 𝒂 𝒆𝒏 𝒕𝒐𝒖𝒕 𝒑𝒐𝒊𝒏𝒕 𝒙 𝒐ù 𝑭𝑿 𝒆𝒔𝒕 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 ∶ 𝑭′𝑿 (𝒙) = 𝒇𝑿 (𝒙).
 𝑭𝑿 (−∞) = 𝐥𝐢𝐦 𝑭𝑿 (𝒙) = 𝟎 𝒆𝒕 𝑭𝑿 (+∞) = 𝐥𝐢𝐦 𝑭𝑿 (𝒙) = 𝟏.
𝒙→−∞ 𝒙→+∞

284 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
 𝑬𝒍𝒍𝒆 𝒄𝒂𝒓𝒂𝒄𝒕é𝒓𝒊𝒔𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝑿 , 𝒂𝒖𝒕𝒓𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒕 ∶ 𝑭𝑿 = 𝑭𝒀 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 𝒍𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔
𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝑿 𝒆𝒕 𝒀 𝒐𝒏𝒕 𝒎ê𝒎𝒆 𝒍𝒐𝒊

𝓐– 𝟒 • 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅𝒆 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 ∶

𝓪 • 𝟎 ≤ 𝑭𝑿 (𝒙) ≤ 𝟏 𝓫 • 𝑷𝑿 (𝑿 ≤ 𝒙) = 𝑷𝑿 (𝑿 < 𝒙) = 𝑷𝑿 (]−∞, 𝒙]) = 𝑭𝑿 (𝒙)
𝓬 • 𝑷𝑿 (𝑿 > 𝒙) = 𝑷𝑿 (]𝒙, +∞[) = 𝟏 − 𝑷𝑿 (𝑿 ≤ 𝒙) = 𝟏 − 𝑭𝑿 (𝒙)
𝓭 • 𝑷(𝑿 ∈ ]𝒙𝟏 , 𝒙𝟐 ]) = 𝑷(𝑿 ∈ [𝒙𝟏 , 𝒙𝟐 ]) = 𝑷(𝑿 ∈ [𝒙𝟏 , 𝒙𝟐 [) = 𝑷(𝑿 ∈ ]𝒙𝟏 , 𝒙𝟐 [) = 𝑭𝑿 (𝒙𝟐 ) − 𝑭𝑿 (𝒙𝟏 )
𝒙𝟐
= ∫ 𝒇𝑿 (𝒕)𝒅𝒕
𝒙𝟏

𝓐– 𝟓 • 𝑳𝒆𝒔 𝒄𝒂𝒓𝒂𝒄𝒕é𝒓𝒊𝒔𝒕𝒊𝒒𝒖𝒆𝒔 𝒅𝒆 𝒕𝒆𝒏𝒅𝒂𝒏𝒄𝒆 𝒄𝒆𝒏𝒕𝒓𝒂𝒍𝒆 ∶

𝑶𝒏 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒒𝒖𝒂𝒏𝒕𝒊𝒍𝒆 𝒐𝒖 𝒇𝒓𝒂𝒄𝒕𝒊𝒍𝒆 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝜶 , (𝟎 ≤ 𝜶 ≤ 𝟏) 𝒅′ 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝑿 𝒅𝒐𝒏𝒕 𝒍𝒂
𝒙𝜶
𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒆𝒔𝒕 𝑭𝑿 (𝒙) , 𝒍𝒂 𝒗𝒂𝒍𝒆𝒖𝒓 𝒙𝜶 𝒕𝒆𝒍𝒍𝒆 𝒒𝒖𝒆 𝑭𝑿 (𝒙𝜶 ) = 𝑷𝑿 (𝑿 ≤ 𝒙𝜶 ) = ∫ 𝒇𝑿 (𝒕)𝒅𝒕 = 𝜶
−∞

𝒙𝜶 𝒔′ 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒒𝒖𝒂𝒏𝒕𝒊𝒍𝒆 𝒐𝒖 𝒇𝒓𝒂𝒄𝒕𝒊𝒍𝒆 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝜶

𝟏
𝓪 • 𝑳𝒂 𝒎é𝒅𝒊𝒂𝒏𝒆 ∶ 𝑳𝒂 𝒎é𝒅𝒊𝒂𝒏𝒆 𝒆𝒔𝒕 𝒍𝒆 𝒒𝒖𝒂𝒏𝒕𝒊𝒍𝒆 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝜶 = 𝒆𝒍𝒍𝒆 𝒔𝒆𝒓𝒂 𝒏𝒐𝒕é𝒆 𝑴𝒆 (𝑿)
𝟐
𝓫 • 𝑳𝒆𝒔 𝒒𝒖𝒂𝒓𝒕𝒊𝒍𝒆𝒔 ∶ 𝑳𝒆𝒔 𝒒𝒖𝒂𝒓𝒕𝒊𝒍𝒆𝒔 , 𝒏𝒐𝒕é𝒔 𝑸𝒊 (𝑿) (𝒓𝒆𝒔𝒑𝒆𝒄𝒕𝒊𝒗𝒆𝒎𝒆𝒏𝒕 𝒊 = 𝟏, 𝟐, 𝟑)
𝒄𝒐𝒓𝒓𝒆𝒔𝒑𝒐𝒏𝒅𝒆𝒏𝒕 𝒂𝒖𝒙 𝒒𝒖𝒂𝒏𝒕𝒊𝒍𝒆𝒔 𝒅′ 𝒐𝒓𝒅𝒓𝒆 (𝜶 = 𝟐𝟓% , 𝟓𝟎% , 𝟕𝟓%) . 𝑵𝒐𝒕𝒐𝒏𝒔 𝒒𝒖𝒆 𝑸𝟐 (𝑿) = 𝑴𝒆 (𝑿)
𝑳’𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆 𝒊𝒏𝒕𝒆𝒓– 𝒒𝒖𝒂𝒓𝒕𝒊𝒍𝒆 𝒆𝒔𝒕 𝒍’𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆 [𝑸𝟏 , 𝑸𝟑 ] . 𝑫𝒆 𝒎ê𝒎𝒆, 𝒐𝒏 𝒂 𝑷𝑿 (𝑿 ∈ [𝑸𝟏 , 𝑸𝟑 ]) = 𝟎, 𝟓
𝓬 • 𝑳𝒆𝒔 𝒅é𝒄𝒊𝒍𝒆𝒔 ∶ 𝑳𝒆 𝒌‑𝒊è𝒎𝒆 𝒅é𝒄𝒊𝒍𝒆 (𝒌 = 𝟏, 𝟐, … , 𝟗) 𝒆𝒔𝒕 𝒍𝒆 𝒒𝒖𝒂𝒓𝒕𝒊𝒍𝒆 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝒌⁄𝟏𝟎
𝓭 • 𝑳𝒆𝒔 𝒄𝒆𝒏𝒕𝒊𝒍𝒆𝒔 ∶ 𝑳𝒆 𝒌‑𝒊è𝒎𝒆 𝒄𝒆𝒏𝒕𝒊𝒍𝒆 (𝒌 = 𝟏, 𝟐, … , 𝟗𝟗) 𝒆𝒔𝒕 𝒍𝒆 𝒒𝒖𝒂𝒓𝒕𝒊𝒍𝒆 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝒌⁄𝟏𝟎𝟎
𝓮 • 𝑳𝒆 𝒎𝒐𝒅𝒆 ∶ 𝑴𝒐 𝒆𝒔𝒕 𝒍𝒂 𝒗𝒂𝒍𝒆𝒖𝒓 𝒅𝒆 𝑿 𝒒𝒖𝒊 𝒎𝒂𝒙𝒊𝒎𝒊𝒔𝒆 𝒇𝑿 (𝒕)

𝓐– 𝟔 • 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶
𝑺𝒐𝒊𝒕 𝑿 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒗é𝒓𝒊𝒇𝒊𝒂𝒏𝒕 ∶ 𝒙𝒇𝑿 (𝒙) 𝒊𝒏𝒕é𝒈𝒓𝒂𝒃𝒍𝒆 𝒔𝒖𝒓 𝑿(𝛀) 𝒆𝒕 𝒕𝒆𝒍 𝒒𝒖𝒆 ∶
+∞
∫ 𝒙𝒇𝑿 (𝒙)𝒅𝒙 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 . 𝑶𝒏 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 𝒅𝒆 𝑿 (𝒐𝒖 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒐𝒖 𝒑𝒓𝒆𝒎𝒊𝒆𝒓
−∞
+∞
𝒎𝒐𝒎𝒆𝒏𝒕 𝒏𝒐𝒏 𝒄𝒆𝒏𝒕𝒓é)𝒍𝒆 𝒓é𝒆𝒍 𝑬(𝑿) 𝒅é𝒇𝒊𝒏𝒊 𝒑𝒂𝒓 ∶ 𝑬(𝑿) = ∫ 𝒙𝒇𝑿 (𝒙)𝒅𝒙
−∞

𝓐– 𝟕 • 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒅’𝒖𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅’𝒖𝒏𝒆 𝒗. 𝒂. ∶

𝑺𝒐𝒊𝒆𝒏𝒕 𝑿 𝒖𝒏𝒆 𝒗. 𝒂. 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒆𝒕 𝝋 𝒖𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒏𝒖𝒎é𝒓𝒊𝒒𝒖𝒆 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒔𝒖𝒓 𝒖𝒏 𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆
+∞
𝒄𝒐𝒏𝒕𝒊𝒆𝒏𝒕 𝑿(𝛀) . 𝑺𝒊 𝑬[𝝋(𝑿)] 𝒆𝒙𝒊𝒔𝒕𝒆 , 𝒂𝒍𝒐𝒓𝒔 ∶ 𝑬[𝝋(𝑿)] = ∫ 𝝋(𝒙)𝒇𝑿 (𝒙)𝒅𝒙
−∞

𝑹𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒔 ∶ 𝑺𝒊 𝑿 𝒆𝒕 𝒀 𝒐𝒏𝒕 𝒎ê𝒎𝒆 𝒍𝒐𝒊 , 𝒊𝒍 𝒆𝒔𝒕 𝒄𝒍𝒂𝒊𝒓 𝒒𝒖𝒆 𝑬(𝑿) = 𝑬(𝒀) . 𝑳𝒂 𝒓é𝒄𝒊𝒑𝒓𝒐𝒒𝒖𝒆 𝒆𝒔𝒕 𝒇𝒂𝒖𝒔𝒔𝒆.

285 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝓐– 𝟖 • 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅𝒆 𝒍′ 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆:
𝓪 • 𝑬(𝒂𝑿 + 𝒃) = 𝒂𝑬(𝑿) + 𝒃 , 𝒆𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 𝑬(𝒂𝑿) = 𝒂𝑬(𝑿) , 𝒐ù 𝒂 𝒆𝒕 𝒃 𝒅𝒆𝒖𝒙 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆𝒔
𝓫 • 𝑬(𝒄) = 𝒄 , 𝒐ù 𝒄 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆. 𝑬𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 𝑬[𝑬(𝑿)] = 𝑬(𝑿) , 𝒑𝒖𝒊𝒔𝒒𝒖𝒆 𝑬(𝑿) 𝒏′ 𝒆𝒔𝒕
𝒑𝒂𝒔 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆
𝓬 • 𝑷𝒐𝒖𝒓 𝒕𝒐𝒖𝒕𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝑿 𝒆𝒕 𝒀 𝒂𝒚𝒂𝒏𝒕 𝒄𝒉𝒂𝒄𝒖𝒏𝒆 𝒖𝒏𝒆 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 , 𝒐𝒏 𝒂 ∶
𝑬(𝑿 + 𝒀) = 𝑬(𝑿) + 𝑬(𝒀) 𝒆𝒕 𝑬(𝑿 − 𝒀) = 𝑬(𝑿) − 𝑬(𝒀)
 𝑮é𝒏é𝒓𝒂𝒍𝒊𝒔𝒂𝒕𝒊𝒐𝒏: 𝑷𝒐𝒖𝒓 𝒕𝒐𝒖𝒕𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. (𝑿𝒊 )𝒊=𝟏,𝟐,…,𝒏 𝒂𝒚𝒂𝒏𝒕 𝒄𝒉𝒂𝒄𝒖𝒏𝒆 𝒖𝒏𝒆 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆,
𝒏 𝒏

𝒐𝒏 𝒂 ∶ 𝑬 (∑ 𝑿𝒊 ) = ∑ 𝑬(𝑿𝒊 )
𝒊=𝟏 𝒊=𝟏

𝓭 • 𝑷𝒐𝒔𝒊𝒕𝒊𝒗𝒊𝒕é 𝒅𝒆 𝒍’𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 ∶
 𝑺𝒊 𝑬(𝑿) 𝒆𝒙𝒊𝒔𝒕𝒆 𝒆𝒕 𝒔𝒊 𝑿 ≥ 𝟎 , 𝒂𝒍𝒐𝒓𝒔 𝑬(𝑿) ≥ 𝟎
 𝑺𝒊 𝑿 𝒆𝒕 𝒀 𝒐𝒏𝒕 𝒖𝒏𝒆 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒆𝒕 𝒔𝒊 𝑿 ≤ 𝒀 , 𝒂𝒍𝒐𝒓𝒔 𝑬(𝑿) ≤ 𝑬(𝒀)
𝓮 • 𝑰𝒏é𝒈𝒂𝒍𝒊𝒕é 𝒅𝒆 𝑱𝒆𝒏𝒔𝒆𝒏 ∶ 𝑺𝒐𝒊𝒕 𝝋 𝒖𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒄𝒐𝒏𝒗𝒆𝒙𝒆 𝒅𝒆 ℝ 𝒗𝒆𝒓𝒔 𝒍𝒖𝒊 𝒎ê𝒎𝒆 𝒆𝒕 𝑿 𝒖𝒏𝒆 𝒗. 𝒂.
𝒕𝒆𝒍𝒍𝒆 𝒒𝒖𝒆 𝑬[𝝋(𝑿)] 𝒆𝒙𝒊𝒔𝒕𝒆 . 𝑶𝒏 𝒂 𝒂𝒍𝒐𝒓𝒔 ∶ 𝝋[𝑬(𝑿)] ≤ 𝑬[𝝋(𝑿)]
 𝑹𝒂𝒑𝒑𝒆𝒍 ∶ 𝑼𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝝋 𝒅𝒆 ℝ 𝒗𝒆𝒓𝒔 𝒍𝒖𝒊 𝒎ê𝒎𝒆 𝒆𝒔𝒕 𝒅𝒊𝒕𝒆 𝒄𝒐𝒏𝒗𝒆𝒙𝒆 𝒔𝒊, 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕 𝒄𝒐𝒖𝒑𝒍𝒆
(𝒙, 𝒚) 𝒅𝒆 ℝ𝟐 𝒆𝒕 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕 𝝀 ∈ [𝟎, 𝟏] , 𝒐𝒏 𝒂 ∶ 𝝋[𝝀𝒙 + (𝟏 − 𝝀)𝒚] ≤ 𝝀𝝋(𝒙) + (𝟏 − 𝝀)𝝋(𝒚)
𝑵𝒐𝒕𝒐𝒏𝒔 , 𝒆𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 , 𝒒𝒖′ 𝒖𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝝋 𝒅𝒆𝒖𝒙 𝒇𝒐𝒊𝒔 𝒅é𝒓𝒊𝒗𝒂𝒃𝒍𝒆𝒔 𝒅𝒐𝒏𝒕 𝒍𝒂 𝒅é𝒓𝒊𝒗é𝒆 𝒔𝒆𝒄𝒐𝒏𝒅𝒆
𝒆𝒔𝒕 𝒑𝒐𝒔𝒊𝒕𝒊𝒗𝒆 (𝝋′′ (𝒙) ≥ 𝟎) 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒄𝒐𝒏𝒗𝒆𝒙𝒆.
𝓯 • 𝑺𝒊 𝑬(𝑿) 𝒆𝒙𝒊𝒔𝒕𝒆 , 𝒂𝒍𝒐𝒓𝒔 |𝑬(𝑿)| ≤ 𝑬(|𝑿|) < +∞
𝓰 • 𝑷𝒐𝒖𝒓 𝒕𝒐𝒖𝒕𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. (𝑿𝒊 )𝒊=𝟏,𝟐,…,𝒏 𝒊𝒅𝒆𝒏𝒕𝒊𝒒𝒖𝒆𝒔 𝒂𝒚𝒂𝒏𝒕 𝒄𝒉𝒂𝒄𝒖𝒏𝒆 𝒖𝒏𝒆 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆

é𝒈𝒂𝒍𝒆 à 𝒎 , 𝒐𝒏 𝒂 ∶ 𝑬(𝑿𝟏 ) = 𝑬(𝑿𝟐 ) = ⋯ = 𝑬(𝑿𝒏 ) = 𝒎

𝓐– 𝟗 • 𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶

𝑺𝒐𝒊𝒕 𝑿 𝒖𝒏𝒆 𝒗. 𝒂. 𝒂𝒚𝒂𝒏𝒕 𝒖𝒏 𝒎𝒐𝒎𝒆𝒏𝒕 𝒏𝒐𝒏 𝒄𝒆𝒏𝒕𝒓é 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝟐 (𝑬(𝑿𝟐 )) .

𝑶𝒏 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒓𝒆𝒔𝒑𝒆𝒄𝒕𝒊𝒗𝒆𝒎𝒆𝒏𝒕 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒅𝒆 𝑿 𝒆𝒕 é𝒄𝒂𝒓𝒕 𝒕𝒚𝒑𝒆 𝒅𝒆 𝑿 𝒍𝒆𝒔 𝒒𝒖𝒂𝒏𝒕𝒊𝒕é𝒔 ∶

+∞
𝟐 𝟐
𝑽𝒂𝒓(𝑿) = 𝑬 [(𝑿 − 𝑬(𝑿)) ] = ∫ (𝒙 − 𝑬(𝑿)) 𝒇𝑿 (𝒙)𝒅𝒙 , 𝝈(𝑿) = √𝑽𝒂𝒓(𝑿)
−∞

𝑳𝒂 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒐𝒖 𝒍𝒆 𝒅𝒆𝒖𝒙𝒊è𝒎𝒆 𝒎𝒐𝒎𝒆𝒏𝒕 𝒄𝒆𝒏𝒕𝒓é 𝒅𝒆 𝑿 𝒆𝒔𝒕 𝒍𝒂 𝒒𝒖𝒂𝒏𝒕𝒊𝒕é 𝒒𝒖𝒊 𝒏𝒐𝒖𝒔 𝒑𝒆𝒓𝒎𝒆𝒕 𝒅𝒆

𝒎𝒆𝒔𝒖𝒓𝒆𝒓 𝒍𝒂 𝒅𝒊𝒔𝒑𝒆𝒓𝒔𝒊𝒐𝒏 𝒂𝒖𝒕𝒐𝒖𝒓 𝒅𝒆 𝒍𝒂 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝑬(𝑿)

𝓐– 𝟏𝟎 • 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅𝒆 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶
𝓪 • ∀ 𝒂 ∈ ℝ , ∀ 𝒃 ∈ ℝ , 𝑽𝒂𝒓(𝒂𝑿 + 𝒃) = 𝒂𝟐 𝑽𝒂𝒓(𝑿) 𝓫 • 𝑽𝒂𝒓(𝒄) = 𝟎 𝒐ù 𝒄 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆.

286 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝑬𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 ∶ 𝑽𝒂𝒓(𝑿) = 𝟎 ⇔ 𝑿 = 𝑬(𝑿) (𝒑. 𝒔. ) ⇔ 𝑿 𝒆𝒔𝒕 𝒑𝒓𝒆𝒔𝒒𝒖𝒆 𝒔û𝒓𝒆𝒎𝒆𝒏𝒕 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆.
+∞
𝟐 𝟐
𝓬 • 𝑭𝒐𝒓𝒎𝒖𝒍𝒆 𝒅𝒆 𝑲𝒐𝒆𝒏𝒊𝒈 ∶ 𝑽𝒂𝒓(𝑿) = 𝑬(𝑿𝟐 ) − (𝑬(𝑿)) = ∫ 𝒙𝟐 𝒇𝑿 (𝒙)𝒅𝒙 − (𝑬(𝑿))
−∞
𝟐
𝓭 • 𝑽𝒂𝒓(𝑿) ≥ 𝟎 𝓮 • 𝑽𝒂𝒓(−𝑿) = 𝑽𝒂𝒓(𝑿) 𝓯 • 𝑽𝒂𝒓(𝑿) = 𝑬 [(𝑿 − 𝑬(𝑿)) ] = 𝐢𝐧𝐟 𝑬[(𝑿 − 𝒂)𝟐 ]
𝒂∈ℝ

𝑿 − 𝑬(𝑿)
𝓰 • 𝑻𝒆𝒓𝒎𝒊𝒏𝒐𝒍𝒐𝒈𝒊𝒆 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊è𝒓𝒆 ∶ 𝑺𝒐𝒊𝒕 𝑿 𝒖𝒏𝒆 𝒗. 𝒂. 𝑳𝒆𝒔 𝒗. 𝒂. 𝑿𝒄 = 𝑿 − 𝑬(𝑿) 𝒆𝒕 𝑿∗ =
𝝈(𝑿)
𝒔𝒐𝒏𝒕 𝒓𝒆𝒔𝒑𝒆𝒄𝒕𝒊𝒗𝒆𝒎𝒆𝒏𝒕 𝒂𝒑𝒑𝒆𝒍é𝒆𝒔 𝒗. 𝒂. 𝒄𝒆𝒏𝒕𝒓é𝒆 𝒆𝒕 𝒗. 𝒂. 𝒄𝒆𝒏𝒕𝒓é𝒆 𝒓é𝒅𝒖𝒊𝒕𝒆 𝒂𝒔𝒔𝒐𝒄𝒊é𝒆𝒔 à 𝑿 .
𝑬(𝑿𝒄 ) = 𝟎 𝑬(𝑿∗ ) = 𝟎
𝑶𝒏 𝒂 ∶ { 𝒆𝒕 {
𝑽𝒂𝒓(𝑿𝒄 ) = 𝑽𝒂𝒓(𝑿) = 𝝈𝟐 𝑽𝒂𝒓(𝑿∗ ) = 𝟏
𝓱 • 𝑷𝒐𝒖𝒓 𝒕𝒐𝒖𝒕𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. (𝑿𝒊 )𝒊=𝟏,𝟐,…,𝒏 𝒊𝒅𝒆𝒏𝒕𝒊𝒒𝒖𝒆𝒔 𝒂𝒚𝒂𝒏𝒕 𝒄𝒉𝒂𝒄𝒖𝒏𝒆 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 é𝒈𝒂𝒍𝒆 à

𝝈𝟐 , 𝒐𝒏 𝒂 ∶ 𝑽𝒂𝒓(𝑿𝟏 ) = 𝑽𝒂𝒓(𝑿𝟐 ) = ⋯ = 𝑽𝒂𝒓(𝑿𝒏 ) = 𝝈𝟐

𝓐– 𝟏𝟏 • 𝑳𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒏𝒐𝒏‑𝒄𝒆𝒏𝒕𝒓é𝒔 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝒓 ∶

𝑺𝒐𝒊𝒕 𝒓 ∈ ℕ∗ . 𝑶𝒏 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒎𝒐𝒎𝒆𝒏𝒕‑𝒄𝒆𝒏𝒕𝒓é𝒔 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝒓 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝑿 , 𝒍𝒂 𝒒𝒖𝒂𝒏𝒕𝒊𝒕é ∶
+∞
𝒓)
𝒎𝒓 = 𝑬(𝑿 =∫ 𝒙𝒓 𝒇𝑿 (𝒙)𝒅𝒙 , 𝒍𝒐𝒓𝒔𝒒𝒖′ 𝒆𝒍𝒍𝒆 𝒆𝒔𝒕 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕𝒆
−∞

 𝑺𝒊 𝑿 𝒑𝒐𝒔𝒔è𝒅𝒆 𝒖𝒏 𝒎𝒐𝒎𝒆𝒏𝒕 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝒓 , 𝒂𝒍𝒐𝒓𝒔 𝒆𝒍𝒍𝒆 𝒑𝒐𝒔𝒔è𝒅𝒆 𝒂𝒖𝒔𝒔𝒊 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒅𝒆

𝒕𝒐𝒖𝒕 𝒐𝒓𝒅𝒓𝒆 𝒏 ≤ 𝒓
 𝑺𝒊 𝒓 = 𝟎 ⇒ 𝒎𝟎 = 𝟏  𝑺𝒊 𝒓 = 𝟏 ⇒ 𝒎𝟏 = 𝑬(𝑿)  𝑺𝒊 𝒓 = 𝟐 ⇒ 𝒎𝟐 = 𝑬(𝑿𝟐 )
 ∀ 𝒓 ∈ ℕ∗ , ∀ (𝜽, 𝝀) ∈ ℝ𝟐 ∶
𝒓 𝒓
𝒓)
𝒎𝒓 (𝜽𝑿 + 𝝀) = 𝑬((𝜽𝑿 + 𝝀) = ∑ 𝑪𝒊𝒓 𝜽𝒓−𝒊 𝝀𝒊 𝒎𝒓−𝒊 (𝑿) = ∑ 𝑪𝒊𝒓 𝜽𝒊 𝝀𝒓−𝒊 𝒎𝒊 (𝑿) , 𝒆𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 ∶
𝒊=𝟎 𝒊=𝟎

𝒎𝟏 (𝜽𝑿 + 𝝀) = 𝑬(𝜽𝑿 + 𝝀) = 𝜽𝑬(𝑿) + 𝝀 = 𝜽𝒎𝟏 (𝑿) + 𝝀

𝓐– 𝟏𝟐 • 𝑳𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒄𝒆𝒏𝒕𝒓é𝒔 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝒓 ∶

𝑺𝒐𝒊𝒕 𝒓 ∈ ℕ∗ . 𝑶𝒏 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒎𝒐𝒎𝒆𝒏𝒕 𝒄𝒆𝒏𝒕𝒓é 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝒓 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝑿 , 𝒍𝒂 𝒒𝒖𝒂𝒏𝒕𝒊𝒕é
+∞
𝒓 𝒓
𝝁𝒓 = 𝑬[(𝑿 − 𝑬(𝑿)) ] = ∫ (𝒙 − 𝑬(𝑿)) 𝒇𝑿 (𝒙)𝒅𝒙 , 𝒍𝒐𝒓𝒔𝒒𝒖′ 𝒆𝒍𝒍𝒆 𝒆𝒔𝒕 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕𝒆
−∞

 𝑺𝒊 𝒓 = 𝟎 ⇒ 𝝁𝟎 = 𝟏  𝑺𝒊 𝒓 = 𝟏 ⇒ 𝝁𝟏 = 𝟎
𝟐 𝟐
 𝑺𝒊 𝒓 = 𝟐 ⇒ 𝝁𝟐 = 𝑬 [(𝑿 − 𝑬(𝑿)) ] = 𝑽𝒂𝒓(𝑿) = 𝑬(𝑿𝟐 ) − (𝑬(𝑿)) ⇒ 𝝁𝟐 = 𝒎𝟐 − 𝒎𝟐𝟏

 ∀ 𝒓 ∈ ℕ∗ , ∀ (𝜽, 𝝀) ∈ ℝ𝟐 ∶ 𝝁𝒓 (𝜽𝑿 + 𝝀) = 𝜽𝒓 𝝁𝒓 (𝑿)

 𝑷𝒐𝒖𝒓 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒔𝒚𝒎é𝒕𝒓𝒊𝒒𝒖𝒆 ∶ 𝝁𝟐𝒌+𝟏 = 𝟎 , ∀𝒌 ≥ 𝟎

𝓐– 𝟏𝟑 • 𝑹𝒆𝒍𝒂𝒕𝒊𝒐𝒏𝒔 𝒆𝒏𝒕𝒓𝒆 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒏𝒐𝒏‑𝒄𝒆𝒏𝒕𝒓é𝒔 𝒆𝒕 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒄𝒆𝒏𝒕𝒓é𝒔 ∶

287 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝓪 • 𝑴𝒐𝒎𝒆𝒏𝒕𝒔 𝒄𝒆𝒏𝒕𝒓é𝒔 𝒆𝒏 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒏𝒐𝒏‑𝒄𝒆𝒏𝒕𝒓é𝒔 ∶
𝒓 𝒓

𝝁𝒓 = ∑ 𝑪𝒊𝒓 𝒎𝒓−𝒊 (−𝒎𝟏 )𝒊 = ∑ 𝑪𝒊𝒓 𝒎𝒊 (−𝒎𝟏 )𝒓−𝒊 , 𝒆𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 ∶

𝒊=𝟎 𝒊=𝟎
𝟐
 𝑷𝒐𝒖𝒓 𝒓 = 𝟐 ∶ 𝝁𝟐 = 𝒎𝟐 − 𝒎𝟐𝟏 𝒐𝒖 𝒆𝒏𝒄𝒐𝒓𝒆 , 𝑽𝒂𝒓(𝑿) = 𝑬(𝑿𝟐 ) − (𝑬(𝑿))
 𝑷𝒐𝒖𝒓 𝒓 = 𝟑 ∶ 𝝁𝟑 = 𝒎𝟑 − 𝟑𝒎𝟐 𝒎𝟏 + 𝟐𝒎𝟑𝟏  𝑷𝒐𝒖𝒓 𝒓 = 𝟒 ∶ 𝝁𝟒 = 𝒎𝟒 − 𝟒𝒎𝟑 𝒎𝟏 + 𝟔𝒎𝟐 𝒎𝟐𝟏 − 𝟑𝒎𝟒𝟏
 𝑷𝒐𝒖𝒓 𝒓 = 𝟓 ∶ 𝝁𝟓 = 𝒎𝟓 − 𝟓𝒎𝟒 𝒎𝟏 + 𝟏𝟎𝒎𝟑 𝒎𝟐𝟏 − 𝟏𝟎𝒎𝟐 𝒎𝟑𝟏 + 𝟒𝒎𝟓𝟏
𝓫 • 𝑴𝒐𝒎𝒆𝒏𝒕𝒔 𝒏𝒐𝒏‑𝒄𝒆𝒏𝒕𝒓é𝒔 𝒆𝒏 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒄𝒆𝒏𝒕𝒓é𝒔 ∶
𝒓 𝒓

𝒎𝒓 = ∑ 𝑪𝒊𝒓 𝝁𝒓−𝒊 𝝁𝒊 = ∑ 𝑪𝒊𝒓 𝝁𝒊 𝝁𝒓−𝒊 , 𝒆𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 ∶

𝒊=𝟎 𝒊=𝟎
𝟐
 𝑷𝒐𝒖𝒓 𝒓 = 𝟐 ∶ 𝒎𝟐 = 𝝁𝟐 + 𝝁𝟐𝟏 𝒐𝒖 𝒆𝒏𝒄𝒐𝒓𝒆 , 𝑬(𝑿𝟐 ) = 𝑽𝒂𝒓(𝑿) + (𝑬(𝑿))
 𝑷𝒐𝒖𝒓 𝒓 = 𝟑 ∶ 𝒎𝟑 = 𝝁𝟑 + 𝟑𝝁𝟐 𝝁𝟏 + 𝝁𝟑𝟏  𝑷𝒐𝒖𝒓 𝒓 = 𝟒 ∶ 𝒎𝟒 = 𝝁𝟒 + 𝟒𝝁𝟑 𝝁𝟏 + 𝟔𝝁𝟐 𝝁𝟐𝟏 + 𝝁𝟒𝟏
 𝑷𝒐𝒖𝒓 𝒓 = 𝟓 ∶ 𝒎𝟓 = 𝝁𝟓 + 𝟓𝝁𝟒 𝝁𝟏 + 𝟏𝟎𝝁𝟑 𝝁𝟐𝟏 + 𝟏𝟎𝝁𝟐 𝝁𝟑𝟏 + 𝝁𝟓𝟏
𝓐– 𝟏𝟒 • 𝑳𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒔𝒚𝒎é𝒕𝒓𝒊𝒆:
𝑳𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒔𝒚𝒎é𝒕𝒓𝒊𝒆 𝜸𝟏 𝒆𝒔𝒕 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒑𝒂𝒓 𝒓𝒂𝒑𝒑𝒐𝒓𝒕 à 𝒍′ 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶
𝟑
𝝁𝟑 𝑬 [(𝑿 − 𝑬(𝑿)) ]
′ ′
𝜸𝟏 = 𝟑 = 𝟑 . 𝑪 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒗𝒂𝒍𝒆𝒖𝒓 𝒔𝒂𝒏𝒔 𝒅𝒊𝒎𝒆𝒏𝒔𝒊𝒐𝒏 𝒒𝒖𝒊 𝒏 𝒆𝒔𝒕 𝒑𝒂𝒔 𝒂𝒇𝒇𝒆𝒄𝒕é 𝒑𝒂𝒓
𝝈 𝟐 𝟐
[𝑬 ((𝑿 − 𝑬(𝑿)) )]

𝒖𝒏 𝒄𝒉𝒂𝒏𝒈𝒆𝒎𝒆𝒏𝒕 𝒅′ 𝒐𝒓𝒊𝒈𝒊𝒏𝒆 𝒆𝒕 𝒅’é𝒄𝒉𝒆𝒍𝒍𝒆. 𝑺𝒆𝒍𝒐𝒏 𝒍𝒂 𝒗𝒂𝒍𝒆𝒖𝒓 𝒅𝒖 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒔𝒚𝒎é𝒕𝒓𝒊𝒆, 𝒍𝒂

𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é𝒑𝒐𝒖𝒓 𝒍𝒆𝒔 𝒗. 𝒂. 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆𝒔, 𝒑𝒓𝒆𝒏𝒅 𝒖𝒏𝒆 𝒇𝒐𝒓𝒎𝒆 𝒅𝒊𝒇𝒇é𝒓𝒆𝒏𝒕𝒆 ∶ é𝒕𝒂𝒍é𝒆 à 𝒅𝒓𝒐𝒊𝒕𝒆 (𝜸𝟏 > 𝟎),

𝒔𝒚𝒎é𝒕𝒓𝒊𝒒𝒖𝒆 (𝜸𝟏 = 𝟎) 𝒐𝒖 é𝒕𝒂𝒍é𝒆 à 𝒈𝒂𝒖𝒄𝒉𝒆 (𝜸𝟏 < 𝟎)

𝓐– 𝟏𝟓 • 𝑳𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒑𝒍𝒂𝒕𝒊𝒔𝒔𝒆𝒎𝒆𝒏𝒕 (𝒌𝒖𝒓𝒕𝒐𝒔𝒊𝒔):
𝑳𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒑𝒍𝒂𝒕𝒊𝒔𝒔𝒆𝒎𝒆𝒏𝒕 𝒗𝒊𝒔𝒆 à 𝒔𝒊𝒕𝒖𝒆𝒓 𝒍𝒂 𝒉𝒂𝒖𝒕𝒆𝒖𝒓 𝒅𝒆 𝒍𝒂 𝒄𝒐𝒖𝒓𝒃𝒆 𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅′ 𝒖𝒏𝒆 𝒍𝒐𝒊
𝒑𝒂𝒓 𝒓𝒂𝒑𝒑𝒐𝒓𝒕 à 𝒍𝒂 𝒓é𝒇é𝒓𝒆𝒏𝒄𝒆𝒒𝒖′ 𝒆𝒔𝒕 𝒍𝒂 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 (𝑳𝒐𝒊 𝒅𝒆 𝑳𝒂𝒑𝒍𝒂𝒄𝒆‑𝑮𝒂𝒖𝒔𝒔).
𝟒
𝝁𝟒 𝑬 [(𝑿 − 𝑬(𝑿)) ]
𝑵𝒐𝒕é 𝜸𝟐 , 𝒔𝒂 𝒇𝒐𝒓𝒎𝒖𝒍𝒆 𝒆𝒔𝒕 𝒍𝒂 𝒔𝒖𝒊𝒗𝒂𝒏𝒕𝒆 ∶ 𝜸𝟐 = 𝟒 − 𝟑 = 𝟐 −𝟑
𝝈 𝟐
[𝑬 ((𝑿 − 𝑬(𝑿)) )]

𝑪′ 𝒆𝒔𝒕 𝒖𝒏 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒔𝒂𝒏𝒔 𝒅𝒊𝒎𝒆𝒏𝒔𝒊𝒐𝒏, 𝒊𝒏𝒗𝒂𝒓𝒊𝒂𝒏𝒕 𝒑𝒂𝒓 𝒄𝒉𝒂𝒏𝒈𝒆𝒎𝒆𝒏𝒕 𝒅′ é𝒄𝒉𝒆𝒍𝒍𝒆 𝒆𝒕 𝒅𝒆

𝒅𝒊𝒎𝒆𝒏𝒔𝒊𝒐𝒏. 𝑳𝒂 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆 𝟑 𝒂 é𝒕é 𝒄𝒉𝒐𝒊𝒔𝒊𝒆 𝒅𝒆 𝒎𝒂𝒏𝒊è𝒓𝒆 à 𝒄𝒆 𝒒𝒖𝒆 𝒍𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕
𝒅′ 𝒂𝒑𝒍𝒂𝒕𝒊𝒔𝒔𝒆𝒎𝒆𝒏𝒕 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝒔𝒐𝒊𝒕 é𝒈𝒂𝒍𝒆 à 𝜸𝟐 = 𝟎 𝑺𝒆𝒍𝒐𝒏 𝒍𝒂 𝒗𝒂𝒍𝒆𝒖𝒓 𝒅𝒖 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕
𝒅′ 𝒂𝒑𝒍𝒂𝒕𝒊𝒔𝒔𝒆𝒎𝒆𝒏𝒕, 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒑𝒐𝒖𝒓 𝒍𝒆𝒔 𝒗. 𝒂. 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆𝒔, 𝒑𝒓𝒆𝒏𝒅 𝒖𝒏𝒆 𝒇𝒐𝒓𝒎𝒆
𝒅𝒊𝒇𝒇é𝒓𝒆𝒏𝒕𝒆 ∶

288 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
 𝜸𝟐 > 𝟎 𝑶𝒏 𝒑𝒂𝒓𝒍𝒆 𝒅𝒆 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒍𝒆𝒑𝒕𝒐𝒌𝒖𝒓𝒕𝒊𝒒𝒖𝒆 (𝒐𝒖 𝒍𝒆𝒑𝒕𝒐𝒄𝒖𝒓𝒕𝒊𝒒𝒖𝒆). 𝑳𝒆𝒔 é𝒄𝒉𝒂𝒏𝒕𝒊𝒍𝒍𝒐𝒏𝒔
𝒂𝒚𝒂𝒏𝒕 𝒅𝒆𝒔 𝒒𝒖𝒆𝒖𝒆𝒔 𝒑𝒍𝒖𝒔 é𝒑𝒂𝒊𝒔𝒔𝒆𝒔 𝒒𝒖𝒆 𝒍𝒂 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝒂𝒖𝒙 𝒆𝒙𝒕𝒓é𝒎𝒊𝒕é𝒔, 𝒊𝒎𝒑𝒍𝒊𝒒𝒖𝒂𝒏𝒕 𝒅𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓𝒔
𝒂𝒏𝒐𝒓𝒎𝒂𝒍𝒆𝒔 𝒑𝒍𝒖𝒔 𝒇𝒓é𝒒𝒖𝒆𝒏𝒕𝒆𝒔
 𝜸𝟐 = 𝟎 ∶ 𝑶𝒏 𝒑𝒂𝒓𝒍𝒆 𝒅𝒆 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒎é𝒔𝒐𝒌𝒖𝒓𝒕𝒊𝒒𝒖𝒆 (𝒐𝒖 𝒎é𝒔𝒐𝒄𝒖𝒓𝒕𝒊𝒒𝒖𝒆). 𝑳𝒂 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆
𝒆𝒔𝒕 𝒖𝒏 𝒄𝒂𝒔 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 𝒅𝒆 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒎é𝒔𝒐𝒌𝒖𝒓𝒕𝒊𝒒𝒖𝒆 𝒑𝒐𝒖𝒓 𝒍𝒂𝒒𝒖𝒆𝒍𝒍𝒆 𝒍𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅𝒆
𝒅𝒊𝒔𝒔𝒚𝒎é𝒕𝒓𝒊𝒆 𝜸𝟏 = 𝟎
 𝜸𝟐 < 𝟎 𝑶𝒏 𝒑𝒂𝒓𝒍𝒆 𝒅𝒆 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒑𝒍𝒂𝒕𝒚𝒌𝒖𝒓𝒕𝒊𝒒𝒖𝒆 (𝒐𝒖 𝒑𝒍𝒂𝒕𝒚𝒄𝒖𝒓𝒕𝒊𝒒𝒖𝒆).
𝑷𝒐𝒖𝒓 𝒖𝒏𝒆 𝒎ê𝒎𝒆 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆, 𝒍𝒂 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒆𝒔𝒕 𝒓𝒆𝒍𝒂𝒕𝒊𝒗𝒆𝒎𝒆𝒏𝒕 « 𝒂𝒑𝒍𝒂𝒕𝒊𝒆 », 𝒔𝒐𝒏 𝒄𝒆𝒏𝒕𝒓𝒆 𝒆𝒕 𝒔𝒆𝒔
𝒒𝒖𝒆𝒖𝒆𝒔 é𝒕𝒂𝒏𝒕 𝒂𝒑𝒑𝒂𝒖𝒗𝒓𝒊𝒆𝒔 𝒂𝒖 𝒑𝒓𝒐𝒇𝒊𝒕 𝒅𝒆𝒔 𝒇𝒍𝒂𝒏𝒄𝒔.

𝓐– 𝟏𝟔 • 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 (𝒐𝒖 𝒕𝒓𝒂𝒏𝒔𝒇𝒐𝒓𝒎é𝒆 𝒅𝒆 𝑳𝒂𝒑𝒍𝒂𝒄𝒆) ∶

𝑺𝒐𝒊𝒕 𝑿 𝒖𝒏𝒆 𝒗. 𝒂. 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒔𝒖𝒓 𝒖𝒏 𝒆𝒔𝒑𝒂𝒄𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒔é (𝜴, 𝓕, 𝓟). 𝑺𝒐𝒊𝒕 𝑰 𝒖𝒏 𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆 𝒄𝒐𝒏𝒕𝒆𝒏𝒂𝒏𝒕
𝟎 , 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕 𝒓é𝒆𝒍 𝒕 ∈ 𝑰 , 𝒐𝒏 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒅𝒆 𝑿
+∞
𝒄𝒐𝒓𝒓𝒆𝒔𝒑𝒐𝒏𝒅𝒂𝒏𝒕 à 𝒕 ∶ 𝑴𝑿 (𝒕) = 𝑬(𝒆𝒕𝑿 ) = ∫ 𝒆𝒕𝒙 𝒇𝑿 (𝒙)𝒅𝒙 , 𝒍𝒐𝒓𝒔𝒒𝒖′ 𝒆𝒍𝒍𝒆 𝒆𝒔𝒕 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕𝒆
−∞

𝑬𝒍𝒍𝒆 𝒈é𝒏è𝒓𝒆 𝒍𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒏𝒐𝒏– 𝒄𝒆𝒏𝒕𝒓é𝒔 ∶ 𝑴𝑿 (𝟎) = 𝟏 , 𝑴′𝑿 (𝟎) = 𝑬(𝑿), 𝑴′′ 𝟐
𝑿 (𝟎) = 𝑬(𝑿 ) …
(𝒓)
𝑴𝑿 (𝟎) = 𝑬(𝑿𝒓 ) = 𝒎𝒓

𝓐– 𝟏𝟕 • 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅𝒆 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 ∶

𝓪 • 𝑳𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒅′ 𝒖𝒏𝒆 𝒗. 𝒂. 𝒄𝒂𝒓𝒂𝒄𝒕é𝒓𝒊𝒔𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝒄𝒆𝒕𝒕𝒆

𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆. 𝑨𝒖𝒕𝒓𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒕 ∶ (𝑴𝑿 (𝒕) = 𝑴𝒀 (𝒕)) ⇔ (𝑿 𝒆𝒕 𝒀 𝒐𝒏𝒕 𝒍𝒂 𝒎ê𝒎𝒆 𝒍𝒐𝒊)

𝓫 • 𝑴𝒂𝑿+𝒃 (𝒕) = 𝒆𝒃𝒕 𝑴𝑿 (𝒂𝒕)

𝓐– 𝟏𝟖 • 𝑪𝒉𝒂𝒏𝒈𝒆𝒎𝒆𝒏𝒕 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 ∶
𝑺𝒐𝒊𝒆𝒏𝒕 𝑰 𝒆𝒕 𝑱 𝒅𝒆𝒖𝒙 𝒐𝒖𝒗𝒆𝒓𝒕𝒔 𝒅𝒆 ℝ 𝒆𝒕 𝑿 𝒖𝒏𝒆 𝒗. 𝒂. 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 à 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 𝒅𝒂𝒏𝒔 𝑰 𝒆𝒕 𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒇𝑿 .
𝑺𝒐𝒊𝒕 𝝋 𝒖𝒏𝒆 𝒃𝒊𝒋𝒆𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝑰 𝒅𝒂𝒏𝒔 𝑱 = 𝑰𝒎(𝝋) , 𝒅é𝒓𝒊𝒗𝒂𝒃𝒍𝒆 𝒆𝒕 𝒊𝒏𝒗𝒆𝒓𝒔𝒊𝒃𝒍𝒆 . 𝑨𝒍𝒐𝒓𝒔 , 𝒍𝒂 𝒗. 𝒂. 𝒀 = 𝝋(𝑿)

𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒔𝒖𝒓 𝑱 𝒂 𝒑𝒐𝒖𝒓 𝒅𝒆𝒏𝒔𝒊𝒕é ∶ 𝒇𝒀 (𝒚) = |(𝝋−𝟏 )′ (𝒚)| [𝒇𝑿 (𝝋−𝟏 (𝒚))] , ∀𝒚 ∈ 𝑱

𝓑– 𝟏 • 𝑳𝒐𝒊 𝑼𝒏𝒊𝒇𝒐𝒓𝒎𝒆 𝑪𝒐𝒏𝒕𝒊𝒏𝒖𝒆 (𝒐𝒖 𝒔𝒖𝒓 𝒖𝒏 𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒂 [𝒂, 𝒃]) 𝓤[𝒂,𝒃] ∶

𝑶𝒏 𝒅𝒊𝒕 𝒒𝒖′ 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝑿 𝒔𝒖𝒊𝒕 𝒍𝒂 𝒍𝒐𝒊 𝒖𝒏𝒊𝒇𝒐𝒓𝒎𝒆 𝒔𝒖𝒓 𝒍𝒆 𝒔𝒆𝒈𝒎𝒆𝒏𝒕 [𝒂, 𝒃] 𝒆𝒕 𝒐𝒏 𝒏𝒐𝒕𝒆
𝑿 ↝ 𝓤[𝒂,𝒃] , (−∞ < 𝒂 < 𝒃 < +∞) 𝒔𝒊 𝒆𝒍𝒍𝒆 𝒂 𝒖𝒏𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒔𝒖𝒓 𝒄𝒆𝒕 𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆 𝒆𝒕

289 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒏𝒖𝒍𝒍𝒆 𝒂𝒊𝒍𝒍𝒆𝒖𝒓𝒔. 𝑶𝒏 𝒂 ∶
𝟏
[𝒂, 𝒃]
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶ 𝒇𝑿 (𝒕) = {𝒃 − 𝒂 , 𝒔𝒊 𝒕 ∈
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
𝟎 , 𝒔𝒊 𝒙 < 𝒂
𝒙−𝒂
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 ∶ 𝑭𝑿 (𝒙) = { , 𝒔𝒊 𝒙 ∈ [𝒂, 𝒃[
𝒃−𝒂
𝟏 , 𝒔𝒊 𝒙 ≥ 𝒃
𝒃+𝒂 (𝒃 − 𝒂)𝟐
 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) =  𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝑿) =
𝟐 𝟏𝟐
𝒃+𝒂
 𝑴é𝒅𝒊𝒂𝒏𝒆 ∶ 𝑴𝒆 (𝑿) =  𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒔𝒚𝒎é𝒕𝒓𝒊𝒆 ∶ 𝜸𝟏 = 𝟎
𝟐
𝟔
 𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒑𝒍𝒂𝒕𝒊𝒔𝒔𝒆𝒎𝒆𝒏𝒕 (𝒌𝒖𝒓𝒕𝒐𝒔𝒊𝒔 𝒏𝒐𝒓𝒎𝒂𝒍𝒊𝒔é) ∶ 𝜸𝟐 = −
𝟓
𝒆𝒕𝒃 − 𝒆𝒕𝒂
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 ∶ 𝑴𝑿 (𝒕) =
𝒕(𝒃 − 𝒂)
 𝑹𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒔 ∶
𝓵([𝒂, 𝒃] ∩ 𝑰)
 𝑺𝒊 𝑿 ↝ 𝓤[𝒂,𝒃] , 𝒂𝒍𝒐𝒓𝒔 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕 𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆 𝑰 𝒅𝒆 ℝ , 𝑷(𝑿 ∈ 𝑰) = , 𝒐ù 𝓵(𝑱) 𝒅é𝒔𝒊𝒈𝒏𝒆
𝓵([𝒂, 𝒃])
𝒍𝒂 𝒍𝒐𝒏𝒈𝒖𝒆𝒖𝒓 𝒅𝒆 𝒍′ 𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆 𝑱. 𝑬𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 𝑷(𝑿 ∈ [𝒂, 𝒃]) = 𝟏
 𝑺𝒊 𝑿 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒓é𝒆𝒍𝒍𝒆 𝒅𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆
𝒆𝒕 𝒔𝒕𝒓𝒊𝒄𝒕𝒆𝒎𝒆𝒏𝒕 𝒄𝒓𝒐𝒊𝒔𝒔𝒂𝒏𝒕𝒆 𝑭 𝒆𝒕 𝒔𝒊 𝑼 ↝ 𝓤[𝟎,𝟏] , 𝒂𝒍𝒐𝒓𝒔 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆

𝒀 = 𝑭−𝟏 (𝑼) 𝒂 𝒎ê𝒎𝒆 𝒍𝒐𝒊 𝒒𝒖𝒆 𝑿

𝑿−𝒂
 𝑺𝒊 𝑿 ↝ 𝓤[𝒂,𝒃] , 𝒂𝒍𝒐𝒓𝒔 𝒀 = ↝ 𝓤[𝟎,𝟏]
𝒃−𝒂
𝓑– 𝟐 • 𝑳𝒐𝒊 𝑬𝒙𝒑𝒐𝒏𝒆𝒏𝒕𝒊𝒆𝒍𝒍𝒆 𝓔(𝝀) ∶
𝑺𝒐𝒊𝒕 𝝀 𝒖𝒏 𝒓é𝒆𝒍 𝒔𝒕𝒓𝒊𝒄𝒕𝒆𝒎𝒆𝒏𝒕 𝒑𝒐𝒔𝒊𝒕𝒊𝒇 . 𝑼𝒏𝒆 𝒗. 𝒂. 𝑿 à 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 𝒅𝒂𝒏𝒔 ℝ+ 𝒆𝒔𝒕 𝒅𝒊𝒕𝒆 𝒅𝒆 𝒍𝒐𝒊
𝒆𝒙𝒑𝒐𝒏𝒆𝒏𝒕𝒊𝒆𝒍𝒍𝒆 𝒅𝒆 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆 𝝀 > 𝟎 , 𝒔𝒊 𝒆𝒍𝒍𝒆 𝒂𝒅𝒎𝒆𝒕 𝒖𝒏𝒆 ∶
𝝀𝒆−𝝀𝒙 , 𝒔𝒊 𝒙 ∈ [𝟎, +∞[
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶ 𝒇𝑿 (𝒙) = {
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
𝟎 , 𝒔𝒊 𝒙 < 𝟎
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 ∶ 𝑭𝑿 (𝒙) = {
𝟏 − 𝒆−𝝀𝒙 , 𝒔𝒊 𝒙 ≥ 𝟎
𝟏 𝟏
 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) =  𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝑿) =
𝝀 𝝀𝟐
 𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒔𝒚𝒎é𝒕𝒓𝒊𝒆 ∶ 𝜸𝟏 = 𝟐
 𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒑𝒍𝒂𝒕𝒊𝒔𝒔𝒆𝒎𝒆𝒏𝒕 (𝒌𝒖𝒓𝒕𝒐𝒔𝒊𝒔) ∶ 𝜸𝟐 = 𝟔

290 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒕 −𝟏
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 ∶ 𝑴𝑿 (𝒕) = (𝟏 − )
𝝀
𝑿𝟏 ↝ 𝓔(𝝀𝟏 )
 𝑺𝒕𝒂𝒃𝒊𝒍𝒊𝒕é ∶ 𝑺𝒊 𝑿𝟏 𝒆𝒕 𝑿𝟐 𝒔𝒐𝒏𝒕 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒕𝒆𝒍𝒍𝒆𝒔 𝒒𝒖𝒆 ∶ { 𝒆𝒕
𝑿𝟐 ↝ 𝓔(𝝀𝟐 )
𝑨𝒍𝒐𝒓𝒔 (𝑿𝟏 + 𝑿𝟐 ) ↝ 𝓔(𝝀𝟏 + 𝝀𝟐 )
 𝑹𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒔 ∶
 𝑰𝒍 𝒆𝒔𝒕 𝒑𝒍𝒖𝒔 𝒄𝒐𝒎𝒎𝒐𝒅𝒆 𝒅′ 𝒖𝒕𝒊𝒍𝒊𝒔𝒆𝒓 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒔𝒖𝒓𝒗𝒊𝒆 ∶
𝟏 , 𝒔𝒊 𝒙 ≤ 𝟎
𝑮𝑿 (𝒙) = 𝑷(𝑿 > 𝒙) = 𝟏 − 𝑭𝑿 (𝒙) = { −𝝀𝒙
𝒆 , 𝒔𝒊 𝒙 > 𝟎
𝑿
 𝑺𝒊 𝑿 ↝ 𝓔(𝟏) ⇒ ∀𝝀 > 𝟎 , ↝ 𝓔(𝝀)
𝝀
𝟏
 𝑺𝒊 𝑿 ↝ 𝓔(𝟏)𝒆𝒕 𝑺𝒊 𝒀 = ⌊𝜽𝑿⌋, 𝜽 > 𝟎 ⇒ 𝒀 ↝ 𝓖 (𝟏 − 𝒆−𝜽 ) , 𝒐𝒊 𝒈é𝒐𝒎é𝒕𝒓𝒊𝒒𝒖𝒆 𝒅𝒆 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆
𝟏
(𝟏 − 𝒆−𝜽 ) , ⌊𝒖⌋ é𝒕𝒂𝒏𝒕 𝒍𝒂 𝒑𝒂𝒓𝒕𝒊𝒆 𝒆𝒏𝒕𝒊è𝒓𝒆 𝒔𝒖𝒑é𝒓𝒊𝒆𝒖𝒓𝒆 𝒅𝒆 𝒖 , 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒑𝒂𝒓 ∶ ⌊𝒖⌋ = 𝐦𝐢𝐧{𝒌 ∈ ℤ/𝒌 ≥ 𝒖}

 𝑼𝒕𝒊𝒍𝒊𝒔𝒂𝒕𝒊𝒐𝒏 ∶ 𝑳𝒆𝒔 𝒍𝒐𝒊𝒔 𝒆𝒙𝒑𝒐𝒏𝒆𝒏𝒕𝒊𝒆𝒍𝒍𝒆𝒔 𝒔𝒐𝒏𝒕 𝒔𝒐𝒖𝒗𝒆𝒏𝒕 𝒖𝒕𝒊𝒍𝒊𝒔é𝒆𝒔 𝒑𝒐𝒖𝒓 𝒎𝒐𝒅é𝒍𝒊𝒔𝒆𝒓 𝒅𝒆𝒔
𝒕𝒆𝒎𝒑𝒔 𝒅’𝒂𝒕𝒕𝒆𝒏𝒕𝒆. 𝑪𝒆𝒕𝒕𝒆 𝒍𝒐𝒊 𝒆𝒔𝒕 𝒕𝒓è𝒔 𝒖𝒕𝒊𝒍𝒊𝒔é𝒆 𝒆𝒏 é𝒕𝒖𝒅𝒆 𝒅𝒆 𝒇𝒊𝒂𝒃𝒊𝒍𝒊𝒕é.
 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅′ 𝒂𝒃𝒔𝒆𝒏𝒄𝒆 𝒅𝒆 𝒎é𝒎𝒐𝒊𝒓𝒆 ∶

① 𝑺𝒊 𝒍𝒂 𝒗. 𝒂. 𝑿 ↝ 𝓔(𝝀) , 𝒂𝒍𝒐𝒓𝒔 𝒆𝒍𝒍𝒆 𝒗é𝒓𝒊𝒇𝒊𝒆 𝒍𝒂

𝒑𝒓𝒐𝒑𝒓𝒊é𝒕é 𝒅′ 𝒂𝒃𝒔𝒆𝒏𝒄𝒆 𝒅𝒆 𝒎é𝒎𝒐𝒊𝒓𝒆 ∶ 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕 𝒕, 𝒔 ≥ 𝟎
𝑷(𝑿 > 𝒔 + 𝒕|𝑿 > 𝒕) = 𝑷(𝑿 > 𝒔) = 𝑮𝑿 (𝒔 + 𝒕)⁄𝑮𝑿 (𝒕) = 𝑮𝑿 (𝒔)

② 𝑹é𝒄𝒊𝒑𝒓𝒐𝒒𝒖𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 𝒖𝒏𝒆 𝒗. 𝒂. 𝑿 𝒗é𝒓𝒊𝒇𝒊𝒆 ① , 𝒂𝒍𝒐𝒓𝒔 𝑿 ↝ 𝓔(𝝀)

𝓑– 𝟑 • 𝑳𝒐𝒊 𝒅𝒆 𝑳𝒂𝒑𝒍𝒂𝒄𝒆– 𝑮𝒂𝒖𝒔𝒔 𝒐𝒖 𝑳𝒐𝒊 𝑵𝒐𝒓𝒎𝒂𝒍𝒆 𝓝(𝒎, 𝝈𝟐 ) ∶

𝑶𝒏 𝒅𝒊𝒕 𝒒𝒖′ 𝒖𝒏𝒆 𝒗. 𝒂. 𝑿 𝒔𝒖𝒊𝒕 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝒐𝒖 𝒅𝒆 𝑳𝒂𝒑𝒍𝒂𝒄𝒆– 𝑮𝒂𝒖𝒔𝒔 𝒅𝒆 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆𝒔 𝒎 ∈ ℝ
𝒆𝒕 𝝈 ∈ ℝ∗+ 𝒆𝒕 𝒐𝒏 𝒏𝒐𝒕𝒆 𝑿 ↝ 𝓝(𝒎, 𝝈𝟐 ) 𝒔𝒊 𝒆𝒍𝒍𝒆 𝒂𝒅𝒎𝒆𝒕 ∶
𝟏 (𝒙 − 𝒎)𝟐
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶ ∀𝒙 ∈ ℝ , 𝒇𝑿 (𝒙) = 𝐞𝐱𝐩 [− ]
𝝈√𝟐𝝅 𝟐𝝈𝟐

𝑶𝒏 𝒅𝒊𝒕 𝒂𝒖𝒔𝒔𝒊 𝒒𝒖𝒆 𝑿 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒗. 𝒂. 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝒐𝒖 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏𝒏𝒆.

 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) = 𝒎  𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝑿) = 𝝈𝟐
 𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒔𝒚𝒎é𝒕𝒓𝒊𝒆 ∶ 𝜸𝟏 = 𝟎
 𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒑𝒍𝒂𝒕𝒊𝒔𝒔𝒆𝒎𝒆𝒏𝒕 (𝒌𝒖𝒓𝒕𝒐𝒔𝒊𝒔) ∶ 𝜸𝟐 = 𝟎
𝝈𝟐 𝒕𝟐
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 ∶ 𝑴𝑿 (𝒕) = 𝐞𝐱𝐩 (𝒎𝒕 + )
𝟐

291 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 ∶

𝓪 • 𝑺𝒊 𝑿 ↝ 𝓝(𝒎, 𝝈𝟐 ) , 𝒑𝒐𝒖𝒓 𝒂 ∈ ℝ∗ 𝒆𝒕 𝒃 ∈ ℝ , 𝒐𝒏 𝒂 ∶ (𝒂𝑿 + 𝒃) ↝ 𝓝 ((𝒂𝒎 + 𝒃), (𝒂𝟐 𝝈𝟐 ))

𝓫 • 𝑺𝒊 𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 𝒏 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒕𝒆𝒍𝒔 𝒒𝒖𝒆 ∀𝒊 = 𝟏, 𝟐, … , 𝒏 ; 𝑿𝒊 ↝ 𝓝(𝒎𝒊 , 𝝈𝟐𝒊 )

𝒏 𝒏 𝒏 𝒏 𝒏 𝒏

𝒂𝒍𝒐𝒓𝒔 , (∑ 𝑿𝒊 ) ↝ 𝓝 (∑ 𝒎𝒊 , ∑ 𝝈𝟐𝒊 ) 𝒆𝒕 𝒃 + (∑ 𝒂𝒊 𝑿𝒊 ) ↝ 𝓝 ([𝒃 + ∑ 𝒂𝒊 𝒎𝒊 ] , [∑ 𝒂𝟐𝒊 𝝈𝟐𝒊 ])

𝒊=𝟏 𝒊=𝟏 𝒊=𝟏 𝒊=𝟏 𝒊=𝟏 𝒊=𝟏

𝒐ù 𝒂𝒊 ∈ ℝ∗ ∀𝒊 = 𝟏, 𝟐, … , 𝒏 𝒆𝒕 𝒃 ∈ ℝ
𝓬 • 𝑺𝒊 (𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 ) 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒊. 𝒊. 𝒅 𝒅𝒆 𝒗. 𝒂. 𝒅𝒆 𝓝(𝒎, 𝝈𝟐 ) , 𝒂𝒍𝒐𝒓𝒔 ∶
𝒏
𝟏 𝝈 𝟐
̅=
𝑿 ∑ 𝑿𝒊 ↝ 𝓝 (𝒎, ( ) )
𝒏 √𝒏
𝒊=𝟏

𝑷(𝒎 − 𝝈 ≤ 𝒀 ≤ 𝒎 + 𝝈) = 𝟔𝟖, 𝟐𝟔%

𝓭 • {𝑷(𝒎 − 𝟐𝝈 ≤ 𝒀 ≤ 𝒎 + 𝟐𝝈) = 𝟗𝟓, 𝟒𝟒%
𝑷(𝒎 − 𝟑𝝈 ≤ 𝒀 ≤ 𝒎 + 𝟑𝝈) = 𝟗𝟗, 𝟕𝟑%

𝓮 • 𝑳𝒂 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝓝(𝒎, 𝝈𝟐 ) 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒗𝒆𝒓𝒔 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝑫𝒊𝒓𝒂𝒄 𝒂𝒖 𝒑𝒐𝒊𝒏𝒕 𝒎 (𝜹𝒎 ) 𝒍𝒐𝒓𝒔𝒒𝒖𝒆 𝝈 → 𝟎
 𝑨𝒑𝒑𝒓𝒐𝒙𝒊𝒎𝒂𝒕𝒊𝒐𝒏 𝒅’𝒖𝒏𝒆 𝒍𝒐𝒊 𝑩𝒊𝒏𝒐𝒎𝒊𝒂𝒍𝒆 𝒆𝒕 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝑷𝒐𝒊𝒔𝒔𝒐𝒏 𝒑𝒂𝒓 𝒖𝒏𝒆 𝒍𝒐𝒊 𝑵𝒐𝒓𝒎𝒂𝒍𝒆:

① 𝑺𝒐𝒊𝒕 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒃𝒊𝒏𝒐𝒎𝒊𝒂𝒍𝒆 𝓑(𝒏, 𝒑) 𝒅𝒆 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒎 = 𝒏𝒑 𝒆𝒕 𝒅′ é𝒄𝒂𝒓𝒕– 𝒕𝒚𝒑𝒆 𝝈 = √𝒏𝒑(𝟏 − 𝒑)

𝒏 > 𝟑𝟎
𝑶𝒏 𝒑𝒆𝒖𝒕 𝒂𝒑𝒑𝒓𝒐𝒙𝒊𝒎𝒆𝒓 𝓑(𝒏, 𝒑) 𝒑𝒂𝒓 𝒍𝒂 𝒍𝒐𝒊 𝓝(𝒎, 𝝈𝟐 ) 𝒅è𝒔 𝒒𝒖𝒆 ∶ {
𝒏𝒑 > 𝟓 𝒆𝒕 𝒏𝒑(𝟏 − 𝒑) > 𝟓

② 𝑺𝒐𝒊𝒕 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒅𝒆 𝒑𝒐𝒊𝒔𝒔𝒐𝒏 𝓟(𝝀)𝒆 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒎 = 𝝀 𝒆𝒕 𝒅′ é𝒄𝒂𝒓𝒕– 𝒕𝒚𝒑𝒆 𝝈 = √𝝀

𝑶𝒏 𝒑𝒆𝒖𝒕 𝒂𝒑𝒑𝒓𝒐𝒙𝒊𝒎𝒆𝒓 𝓟(𝝀) 𝒑𝒂𝒓 𝒍𝒂 𝒍𝒐𝒊 𝓝(𝒎, 𝝈𝟐 ) 𝒅è𝒔 𝒒𝒖𝒆 ∶ 𝝀 > 𝟏𝟓
 𝑼𝒕𝒊𝒍𝒊𝒔𝒂𝒕𝒊𝒐𝒏 ∶ 𝑼𝒏𝒆 𝒗. 𝒂. 𝑿 𝒔𝒖𝒊𝒕 𝒍𝒂 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 (𝑳𝒐𝒊 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏𝒏𝒆, 𝑳𝒐𝒊 𝒅𝒆
𝑳𝒂𝒑𝒍𝒂𝒄𝒆– 𝑮𝒂𝒖𝒔𝒔), 𝒍𝒐𝒓𝒔𝒒𝒖′ 𝒆𝒍𝒍𝒆 𝒆𝒔𝒕 𝒇𝒐𝒓𝒎é𝒆à 𝒑𝒂𝒓𝒕𝒊𝒓 𝒅′𝒖𝒏 𝒈𝒓𝒂𝒏𝒅 𝒏𝒐𝒎𝒃𝒓𝒆 𝒅𝒆 𝒇𝒂𝒄𝒕𝒆𝒖𝒓𝒔
𝒔′ 𝒂𝒅𝒅𝒊𝒕𝒊𝒐𝒏𝒏𝒂𝒏𝒕 𝒍𝒆𝒔 𝒖𝒏𝒔 𝒂𝒖𝒙 𝒂𝒖𝒕𝒓𝒆𝒔, 𝒂𝒖𝒄𝒖𝒏𝒆 𝒏𝒆 𝒋𝒐𝒖𝒂𝒏𝒕 𝒖𝒏 𝒓ô𝒍𝒆 𝒑𝒓é𝒅𝒐𝒎𝒊𝒏𝒂𝒏𝒕, 𝒆𝒕 𝒂𝒈𝒊𝒔𝒔𝒂𝒏𝒕
𝒅𝒆 𝒎𝒂𝒏𝒊è𝒓𝒆 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆.
𝑪𝒐𝒏𝒔𝒊𝒅é𝒓é𝒆 𝒑𝒆𝒏𝒅𝒂𝒏𝒕 𝒍𝒐𝒏𝒈𝒕𝒆𝒎𝒑𝒔 𝒄𝒐𝒎𝒎𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒖𝒏𝒊𝒗𝒆𝒓𝒔𝒆𝒍𝒍𝒆, 𝒔𝒆𝒔 𝒄𝒂𝒓𝒂𝒄𝒕é𝒓𝒊𝒔𝒕𝒊𝒒𝒖𝒆𝒔 𝒔𝒆
𝒓𝒆𝒕𝒓𝒐𝒖𝒗𝒂𝒏𝒕 𝒅𝒂𝒏𝒔 𝒅𝒆 𝒏𝒐𝒎𝒃𝒓𝒆𝒖𝒙 𝒅𝒐𝒎𝒂𝒊𝒏𝒆𝒔, 𝒅𝒂𝒏𝒔 𝒍𝒆𝒔 𝒔𝒄𝒊𝒆𝒏𝒄𝒆𝒔 𝒅𝒖𝒓𝒆𝒔 𝒎𝒂𝒊𝒔 é𝒈𝒂𝒍𝒆𝒎𝒆𝒏𝒕
𝒅𝒂𝒏𝒔 𝒍𝒆𝒔 𝒔𝒄𝒊𝒆𝒏𝒄𝒆𝒔 𝒉𝒖𝒎𝒂𝒊𝒏𝒆𝒔 𝒆𝒕 𝒔𝒐𝒄𝒊𝒂𝒍𝒆𝒔 (𝒆𝒙. 𝒇𝒍𝒖𝒄𝒕𝒖𝒂𝒕𝒊𝒐𝒏𝒔 é𝒄𝒐𝒏𝒐𝒎𝒊𝒒𝒖𝒆𝒔), 𝒆𝒍𝒍𝒆 𝒂 é𝒕é
𝒅𝒆𝒑𝒖𝒊𝒔 𝒓𝒆𝒄𝒂𝒅𝒓é𝒆.
𝑵é𝒂𝒏𝒎𝒐𝒊𝒏𝒔, 𝒍𝒂 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝒈𝒂𝒓𝒅𝒆 𝒖𝒏𝒆 𝒑𝒍𝒂𝒄𝒆 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊è𝒓𝒆 𝒅𝒆 𝒑𝒂𝒓 𝒔𝒂 𝒔𝒊𝒎𝒑𝒍𝒊𝒄𝒊𝒕é.
𝑬𝒍𝒍𝒆 𝒆𝒔𝒕 à 𝒍𝒂 𝒃𝒂𝒔𝒆 𝒅𝒆 𝒏𝒐𝒎𝒃𝒓𝒆𝒖𝒙 𝒓é𝒔𝒖𝒍𝒕𝒂𝒕𝒔 𝒅𝒂𝒏𝒔 𝒍𝒂 𝒔𝒕𝒂𝒕𝒊𝒔𝒕𝒊𝒒𝒖𝒆 𝒑𝒂𝒓𝒂𝒎é𝒕𝒓𝒊𝒒𝒖𝒆, 𝒆𝒏 𝒕𝒉é𝒐𝒓𝒊𝒆
𝒅𝒆 𝒍′ 𝒆𝒔𝒕𝒊𝒎𝒂𝒕𝒊𝒐𝒏 𝒆𝒕 𝒕𝒉é𝒐𝒓𝒊𝒆 𝒅𝒆𝒔 𝒕𝒆𝒔𝒕𝒔.
𝑫𝒂𝒏𝒔 𝒍𝒂 𝒓é𝒈𝒓𝒆𝒔𝒔𝒊𝒐𝒏 𝒍𝒊𝒏é𝒂𝒊𝒓𝒆 𝒎𝒖𝒍𝒕𝒊𝒑𝒍𝒆, 𝒍𝒆𝒔 𝒍𝒐𝒊𝒔 𝒖𝒕𝒊𝒍𝒊𝒔é𝒆𝒔 𝒑𝒐𝒖𝒓 𝒍′ é𝒗𝒂𝒍𝒖𝒂𝒕𝒊𝒐𝒏 𝒈𝒍𝒐𝒃𝒂𝒍𝒆

292 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒅𝒆 𝒍𝒂 𝒓é𝒈𝒓𝒆𝒔𝒔𝒊𝒐𝒏 𝒆𝒕 𝒍′ é𝒗𝒂𝒍𝒖𝒂𝒕𝒊𝒐𝒏 𝒊𝒏𝒅𝒊𝒗𝒊𝒅𝒖𝒆𝒍𝒍𝒆 𝒅𝒆𝒔 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕𝒔 𝒓𝒆𝒑𝒐𝒔𝒆𝒏𝒕 𝒔𝒖𝒓 𝒍𝒂
𝒏𝒐𝒓𝒎𝒂𝒍𝒊𝒕é 𝒅𝒆𝒔 𝒓é𝒔𝒊𝒅𝒖𝒔.

𝓑– 𝟒 • 𝑳𝒐𝒊 𝑵𝒐𝒓𝒎𝒂𝒍𝒆 𝒄𝒆𝒏𝒕𝒓é𝒆 𝒓é𝒅𝒖𝒊𝒕𝒆 𝒐𝒖 𝑳𝒐𝒊 𝑵𝒐𝒓𝒎𝒂𝒍𝒆 𝒔𝒕𝒂𝒏𝒅𝒂𝒓𝒅 𝓝(𝟎, 𝟏) ∶

𝑿−𝒎
𝑺𝒊 𝑿 ↝ 𝓝(𝒎, 𝝈𝟐 ) , 𝒂𝒍𝒐𝒓𝒔 𝒀 = ↝ 𝓝(𝟎, 𝟏)
𝝈
𝟏 𝒚𝟐
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶ ∀𝒚 ∈ ℝ , 𝒇𝒀 (𝒚) = 𝐞− 𝟐
√𝟐𝝅
 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝒀) = 𝟎  𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝒀) = 𝟏
𝒕𝟐
 𝑴é𝒅𝒊𝒂𝒏𝒆 ∶ 𝑴𝒆 (𝒀) = 𝟎  𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 ∶ 𝑴𝒀 (𝒕) = 𝐞 𝟐
𝒚 𝒖𝟐
𝟏
 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅𝒆 𝒍𝒂 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 ∶ 𝑭𝒀 (𝒚) = 𝚽𝒀 (𝒚) = ∫ 𝐞− 𝟐 𝐝𝐮
−∞ √𝟐𝝅

𝓪 • 𝚽(𝒚) = 𝑷(𝒀 ≤ 𝒚) = 𝑷(𝒀 ≥ −𝒚) = 𝟏 − 𝑷(𝒀 ≤ −𝒚) = 𝟏 − 𝚽(−𝒚)

𝓫 • 𝑷𝒐𝒖𝒓 𝒚 ≥ 𝟒 , 𝚽(𝒚) ≅ 𝟏 𝒆𝒕 𝚽(−𝒚) ≅ 𝟎 𝓬 • 𝑷(|𝒀| ≤ 𝒚) = 𝟐𝚽(𝒚) − 𝟏

𝓑– 𝟓 • 𝑳𝒐𝒊 𝑳𝒐𝒈– 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝓛𝓝(𝒎, 𝝈𝟐 ) ∶

𝑶𝒏 𝒅𝒊𝒕 𝒒𝒖𝒆 𝑿 ↝ 𝓛𝓝(𝒎, 𝝈𝟐 ) , 𝒔𝒊 𝒀 = 𝐥𝐧 𝑿 ↝ 𝓝(𝒎, 𝝈𝟐 ) . 𝑪′ 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒔𝒖𝒓 [𝟎, +∞] .
𝟏 [𝐥𝐧(𝒙) − 𝒎]𝟐
𝐞𝐱𝐩 (− ) , 𝐬𝐢 𝒙 > 𝟎
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶ 𝒇𝑿 (𝒙) = {𝝈𝒙√𝟐𝝅 𝟐𝝈𝟐
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 ∶ ∀𝒙 > 𝟎 , 𝑭𝑿 (𝒙) = 𝑭𝒀 (𝐥𝐧 𝒙)
𝝈𝟐
(𝒎+ )
𝟐
 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) = 𝑬(𝒆𝒀 ) = 𝒆
𝟐 𝟐
 𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝑿) = 𝑽𝒂𝒓(𝒆𝒀 ) = (𝟏 − 𝒆−𝝈 )(𝒆𝟐(𝒎+𝝈 ) )

 𝑼𝒕𝒊𝒍𝒊𝒔𝒂𝒕𝒊𝒐𝒏 ∶ 𝑶𝒏 𝒓𝒆𝒏𝒄𝒐𝒏𝒕𝒓𝒆 𝒇𝒓é𝒒𝒖𝒆𝒎𝒎𝒆𝒏𝒕 𝒄𝒆𝒕𝒕𝒆 𝒍𝒐𝒊 𝒅𝒂𝒏𝒔 𝒍𝒂 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒅𝒆𝒔

𝒔𝒂𝒍𝒂𝒊𝒓𝒆𝒔 𝒐𝒖 𝒅𝒆𝒔 𝒓𝒆𝒗𝒆𝒏𝒖𝒔. 𝑼𝒏 𝒈𝒓𝒂𝒏𝒅 𝒏𝒐𝒎𝒃𝒓𝒆 𝒅𝒆 𝒑𝒆𝒓𝒔𝒐𝒏𝒏𝒆𝒔 𝒅𝒊𝒔𝒑𝒐𝒔𝒆𝒏𝒕 𝒅′ 𝒖𝒏 𝒔𝒂𝒍𝒂𝒊𝒓𝒆
𝒎𝒐𝒅𝒆𝒔𝒕𝒆 (𝒑𝒓𝒐𝒄𝒉𝒆 𝒅𝒖 𝑺𝑴𝑰𝑪 𝒑𝒂𝒓 𝒆𝒙𝒆𝒎𝒑𝒍𝒆), 𝒆𝒕 à 𝒎𝒆𝒔𝒖𝒓𝒆 𝒒𝒖𝒆 𝒍𝒂 𝒗𝒂𝒍𝒆𝒖𝒓 𝒔′ é𝒍è𝒗𝒆, 𝒍𝒂
𝒑𝒓𝒐𝒑𝒐𝒓𝒕𝒊𝒐𝒏 𝒅𝒆𝒔 𝒑𝒆𝒓𝒔𝒐𝒏𝒏𝒆𝒔 𝒄𝒐𝒏𝒄𝒆𝒓𝒏é𝒆𝒔 𝒅𝒊𝒎𝒊𝒏𝒖𝒆 𝒓𝒂𝒑𝒊𝒅𝒆𝒎𝒆𝒏𝒕. 𝑳𝒂 𝒎é𝒅𝒊𝒂𝒏𝒆 𝒆𝒔𝒕 𝒕𝒓è𝒔
𝒊𝒏𝒇é𝒓𝒊𝒆𝒖𝒓𝒆 à 𝒍𝒂 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒑𝒐𝒖𝒓 𝒄𝒆 𝒕𝒚𝒑𝒆 𝒅𝒆 𝒍𝒐𝒊.

𝓑– 𝟔 • 𝑳𝒐𝒊𝒔 𝒅𝒆 𝑪𝒂𝒖𝒄𝒉𝒚 (𝒐𝒖 𝑳𝒐𝒊 𝒅𝒆 𝑳𝒐𝒓𝒆𝒏𝒕𝒛) 𝓒(𝟎 , 𝒂) ∶

𝑼𝒏𝒆 𝒗. 𝒂. à 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 𝒅𝒂𝒏𝒔 ℝ 𝒆𝒔𝒕 𝒅𝒊𝒕𝒆 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝑪𝒂𝒖𝒄𝒉𝒚 𝓒(𝟎 , 𝒂) , 𝒂 > 𝟎 , 𝒔𝒊 𝒆𝒍𝒍𝒆 𝒂𝒅𝒎𝒆𝒕 ∶
𝒂⁄𝝅
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶ ∀𝒙 ∈ ℝ , 𝒇𝑿 (𝒙) =
𝒂𝟐+ 𝒙𝟐
𝟏 𝒙 𝟏
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 ∶ 𝑭𝑿 (𝒙) = 𝐚𝐫𝐜𝐭𝐚𝐧 ( ) +
𝝅 𝒂 𝟐

293 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) 𝒏′𝒆𝒔𝒕 𝒑𝒂𝒔 𝒅é𝒇𝒊𝒏𝒊𝒆
 𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝑿) 𝒏′ 𝒆𝒔𝒕 𝒑𝒂𝒔 𝒅é𝒇𝒊𝒏𝒊𝒆
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 ∶ 𝑴𝑿 (𝒕)𝒏′ 𝒆𝒔𝒕 𝒑𝒂𝒔 𝒅é𝒇𝒊𝒏𝒊𝒆
 𝑹𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒔 ∶ 𝑰𝒍 𝒂𝒓𝒓𝒊𝒗𝒆 𝒑𝒂𝒓𝒇𝒐𝒊𝒔 𝒒𝒖’𝒐𝒏 𝒄𝒐𝒏𝒔𝒊𝒅è𝒓𝒆 𝒅𝒆𝒔 𝒍𝒐𝒊𝒔 𝒅𝒆 𝑪𝒂𝒖𝒄𝒉𝒚 𝒄𝒆𝒏𝒕𝒓é𝒆𝒔
𝒆𝒏 𝒙𝟎 ≠ 𝟎 , 𝓒(𝒙𝟎 , 𝒂) . 𝑫𝒂𝒏𝒔 𝒄𝒆 𝒄𝒂𝒔 , 𝒍𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒆𝒔𝒕 𝒅𝒐𝒏𝒏é𝒆 𝒑𝒂𝒓 ∶
𝒂⁄𝝅
∀𝒙 ∈ ℝ , 𝒇𝑿 (𝒙) =
𝒂𝟐 + (𝒙 − 𝒙𝟎 )𝟐

𝓑– 𝟕 • 𝑳𝒐𝒊 𝒅𝒆 𝑷𝒂𝒓𝒆𝒕𝒐 𝕻(𝒄 , 𝜶):

𝑳𝒂 𝒍𝒐𝒊 𝒅𝒆 𝑷𝒂𝒓𝒆𝒕𝒐 𝒔′ 𝒂𝒑𝒑𝒍𝒊𝒒𝒖𝒆 𝒑𝒐𝒖𝒓 𝒍𝒆𝒔 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏𝒔 𝒕𝒓𝒐𝒏𝒒𝒖é𝒆𝒔.
𝑷𝒓𝒆𝒏𝒐𝒏𝒔 𝒖𝒏 𝒆𝒙𝒆𝒎𝒑𝒍𝒆 𝒅𝒆 𝒍𝒂 𝒗𝒊𝒆 𝒄𝒐𝒖𝒓𝒂𝒏𝒕𝒆, 𝒍𝒂 𝒃𝒐𝒓𝒏𝒆 𝒃𝒂𝒔𝒔𝒆 𝒅𝒖 𝒔𝒂𝒍𝒂𝒊𝒓𝒆 𝒉𝒐𝒓𝒂𝒊𝒓𝒆 𝒆𝒔𝒕
𝒇𝒐𝒓𝒄é𝒎𝒆𝒏𝒕 𝒍𝒆 𝑺𝑴𝑰𝑪, 𝒊𝒍 𝒏𝒆 𝒑𝒆𝒖𝒕 𝒑𝒂𝒔 𝒆𝒏 ê𝒕𝒓𝒆 𝒂𝒖𝒕𝒓𝒆𝒎𝒆𝒏𝒕.
𝑳𝒂 𝒍𝒐𝒊 𝒅𝒆 𝑷𝒂𝒓𝒆𝒕𝒐 𝒑𝒆𝒓𝒎𝒆𝒕 𝒅𝒆 𝒕𝒆𝒏𝒊𝒓 𝒄𝒐𝒎𝒑𝒕𝒆 𝒅𝒆 𝒄𝒆𝒕𝒕𝒆 𝒄𝒐𝒏𝒕𝒓𝒂𝒊𝒏𝒕𝒆 𝒆𝒏 𝒓𝒆𝒔𝒕𝒓𝒆𝒊𝒈𝒏𝒂𝒏𝒕 𝒍𝒆
𝒅𝒐𝒎𝒂𝒊𝒏𝒆 𝒅𝒆 𝒅é𝒇𝒊𝒏𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝑿
𝑳𝒂 𝒍𝒐𝒊 𝒑𝒐𝒔𝒔è𝒅𝒆 𝟐 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆𝒔, 𝜶 > 𝟎 𝒆𝒕 𝒄 𝒒𝒖𝒊 𝒊𝒏𝒕𝒓𝒐𝒅𝒖𝒊𝒕 𝒍𝒂 𝒄𝒐𝒏𝒕𝒓𝒂𝒊𝒏𝒕𝒆 𝒙 > 𝒄 . 𝑳𝒆 𝒅𝒐𝒎𝒂𝒊𝒏𝒆
𝒅𝒆 𝒅é𝒇𝒊𝒏𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝑿 𝒆𝒔𝒕 ]𝒄, +∞[
𝜶 𝒄 𝜶+𝟏
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶ ∀𝒙 ∈ ]𝒄, +∞[ , 𝒇𝑿 (𝒙) = ( )
𝒄 𝒙
𝟎 , 𝒔𝒊 𝒙 < 𝒄
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 ∶ 𝑭𝑿 (𝒙) = {
𝟏 − (𝒄⁄𝒙)𝜶 , 𝒔𝒊 𝒙 ≥ 𝒄
𝜶
 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) = ( ) 𝒄 , 𝒑𝒐𝒖𝒓 𝜶 > 𝟏
𝜶−𝟏
𝜶
𝒆𝒕 𝑬(𝑿𝒓 ) = ( ) 𝒄𝒓 , 𝒑𝒐𝒖𝒓 𝜶 > 𝒓
𝜶−𝒓
𝜶
 𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝑿) = ( ) 𝒄𝟐 , 𝒑𝒐𝒖𝒓 𝜶 > 𝟐
(𝜶 − 𝟏)𝟐 (𝜶 − 𝟐)
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 ∶ 𝑴𝑿 (𝒕) 𝒏′ 𝒆𝒔𝒕 𝒑𝒂𝒔 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒑𝒐𝒖𝒓 𝒕 > 𝟎
 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅′ 𝒖𝒏𝒆 𝒍𝒐𝒊 à 𝒒𝒖𝒆𝒖𝒆 𝒍𝒐𝒏𝒈𝒖𝒆 (𝒐𝒖 𝒍𝒐𝒏𝒈𝒖𝒆 𝒕𝒓𝒂î𝒏𝒆 ) ∶ 𝑳𝒂 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒅𝒆
𝑷𝒂𝒓𝒆𝒕𝒐 𝒆𝒔𝒕 à 𝒒𝒖𝒆𝒖𝒆 𝒍𝒐𝒏𝒈𝒖𝒆 ∶ 𝐥𝐢𝐦 𝑷(𝑿 > 𝒙 + 𝒚|𝑿 > 𝒙) = 𝟏 , 𝒑𝒐𝒖𝒓 𝒚 > 𝟎
𝒙→+∞

𝑷𝒂𝒓 𝒆𝒙𝒆𝒎𝒑𝒍𝒆, 𝒔𝒊 𝑿 𝒆𝒔𝒕 𝒍𝒆 𝒕𝒆𝒎𝒑𝒔 𝒅𝒆 𝒗𝒊𝒆 𝒅′ 𝒖𝒏 𝒄𝒐𝒎𝒑𝒐𝒔𝒂𝒏𝒕, 𝒑𝒍𝒖𝒔 𝒊𝒍 𝒂 𝒗é𝒄𝒖 (𝑿 > 𝒙) 𝒑𝒍𝒖𝒔 𝒊𝒍 𝒂 𝒅𝒆
𝒄𝒉𝒂𝒏𝒄𝒆𝒔 𝒅𝒆 𝒗𝒊𝒗𝒓𝒆 𝒍𝒐𝒏𝒈𝒕𝒆𝒎𝒑𝒔 ∶ 𝒍𝒆 𝒔𝒚𝒔𝒕è𝒎𝒆 𝒓𝒂𝒋𝒆𝒖𝒏𝒊𝒕.
𝑶𝒏 𝒑𝒆𝒖𝒕 𝒑𝒂𝒍𝒍𝒊𝒆𝒓 𝒍′ 𝒊𝒏𝒄𝒐𝒏𝒗é𝒏𝒊𝒆𝒏𝒕 « 𝒍𝒐𝒏𝒈𝒖𝒆 𝒒𝒖𝒆𝒖𝒆 » 𝒅𝒂𝒏𝒔 𝒅′ 𝒂𝒖𝒕𝒓𝒆𝒔 𝒂𝒑𝒑𝒍𝒊𝒄𝒂𝒕𝒊𝒐𝒏𝒔 𝒅𝒆𝒔
𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏𝒔 𝒅𝒆 𝑷𝒂𝒓𝒆𝒕𝒐 𝒕𝒆𝒍𝒍𝒆𝒔 𝒒𝒖𝒆 𝒍𝒂 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒑𝒂𝒓 𝒕𝒂𝒊𝒍𝒍𝒆 𝒅𝒆𝒔 𝒆𝒏𝒕𝒓𝒆𝒑𝒓𝒊𝒔𝒆𝒔 𝒆𝒙𝒑𝒓𝒊𝒎é𝒆
𝒆𝒏 𝒏𝒐𝒎𝒃𝒓𝒆 𝒅′ 𝒆𝒎𝒑𝒍𝒐𝒚é𝒔 𝒐𝒖 𝒆𝒏 𝒄𝒉𝒊𝒇𝒇𝒓𝒆 𝒅′ 𝒂𝒇𝒇𝒂𝒊𝒓𝒆𝒔 𝒐𝒖 𝒅′ 𝒂𝒖𝒕𝒓𝒆𝒔 𝒆𝒏𝒕𝒊𝒕é𝒔 𝒎𝒆𝒔𝒖𝒓𝒂𝒃𝒍𝒆𝒔 𝒑𝒂𝒓

294 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒕𝒂𝒊𝒍𝒍𝒆 𝒅𝒐𝒏𝒕 𝒍𝒂 𝒍𝒊𝒎𝒊𝒕𝒆 𝒕𝒉é𝒐𝒓𝒊𝒒𝒖𝒆 𝒆𝒔𝒕 𝒊𝒏𝒇𝒊𝒏𝒊𝒆 𝒆𝒏 𝒖𝒕𝒊𝒍𝒊𝒔𝒂𝒏𝒕 𝒖𝒏𝒆 é𝒄𝒉𝒆𝒍𝒍𝒆 𝒍𝒐𝒈‑ 𝒍𝒐𝒈 𝒂𝒑𝒓è𝒔
𝒕𝒓𝒂𝒏𝒔𝒇𝒐𝒓𝒎𝒂𝒕𝒊𝒐𝒏𝒔 𝒂𝒑𝒑𝒓𝒐𝒑𝒓𝒊é𝒆𝒔 𝒅𝒆𝒔 𝒅𝒐𝒏𝒏é𝒆𝒔 𝒂𝒏𝒂𝒍𝒚𝒔é𝒆𝒔. 𝑳𝒆 𝒑𝒉é𝒏𝒐𝒎è𝒏𝒆 𝒍𝒐𝒏𝒈𝒖𝒆 𝒒𝒖𝒆𝒖𝒆
𝒆𝒔𝒕 𝒄𝒂𝒖𝒔é 𝒑𝒂𝒓 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒑𝒐𝒖𝒗𝒂𝒏𝒕 𝒂𝒕𝒕𝒆𝒊𝒏𝒅𝒓𝒆 𝒅𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 𝒕𝒓è𝒔 𝒈𝒓𝒂𝒏𝒅𝒆𝒔, 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 𝒑𝒐𝒖𝒓
𝒍𝒆𝒔𝒒𝒖𝒆𝒍𝒍𝒆𝒔 𝒍𝒆 𝒏𝒐𝒎𝒃𝒓𝒆 𝒅′ 𝒐𝒃𝒔𝒆𝒓𝒗𝒂𝒕𝒊𝒐𝒏𝒔 𝒅𝒆𝒗𝒊𝒆𝒏𝒕 𝒕𝒓è𝒔 𝒑𝒆𝒕𝒊𝒕 ; 𝒆𝒏 𝒓𝒆𝒗𝒂𝒏𝒄𝒉𝒆 𝒍𝒆 𝒏𝒐𝒎𝒃𝒓𝒆
𝒅′ 𝒐𝒃𝒔𝒆𝒓𝒗𝒂𝒕𝒊𝒐𝒏𝒔 𝒑𝒐𝒖𝒓 𝒍𝒆𝒔 𝒑𝒆𝒕𝒊𝒕𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 𝒅𝒆 𝒍𝒂 𝒕𝒂𝒊𝒍𝒍𝒆 𝒂𝒏𝒂𝒍𝒚𝒔é𝒆 𝒔𝒐𝒏𝒕 𝒔𝒐𝒖𝒗𝒆𝒏𝒕 𝒕𝒓è𝒔 é𝒍𝒆𝒗é𝒆𝒔.
𝑫𝒂𝒏𝒔 𝒄𝒆 𝒄𝒂𝒔, 𝒐𝒏 𝒂 𝒍𝒆 𝒑𝒉é𝒏𝒐𝒎è𝒏𝒆 𝒔𝒚𝒎é𝒕𝒓𝒊𝒒𝒖𝒆 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒏𝒈𝒖𝒆 𝒒𝒖𝒆𝒖𝒆 ∶
𝒍𝒆 𝒍𝒐𝒏𝒈 𝒑𝒊𝒄 𝒊𝒏𝒊𝒕𝒊𝒂𝒍. 𝑫𝒂𝒏𝒔 𝒍𝒆 𝒄𝒂𝒔 𝒅𝒆 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏𝒔 𝒅𝒆 𝑷𝒂𝒓𝒆𝒕𝒐, 𝒍𝒆 𝒑𝒂𝒔𝒔𝒂𝒈𝒆 𝒆𝒏 𝒄𝒐𝒐𝒓𝒅𝒐𝒏𝒏é𝒆𝒔
𝒍𝒐𝒈 𝒍𝒐𝒈 𝒕𝒓𝒂𝒏𝒔𝒇𝒐𝒓𝒎𝒆 𝒆𝒏 𝒍𝒊𝒈𝒏𝒆 𝒅𝒓𝒐𝒊𝒕𝒆 𝒍𝒂 𝒄𝒐𝒖𝒓𝒃𝒆 𝒅𝒐𝒏𝒕 𝒍𝒂 𝒇𝒐𝒓𝒎𝒆 𝒐𝒓𝒊𝒈𝒊𝒏𝒂𝒍𝒆 𝒆𝒔𝒕 𝒖𝒏𝒆
𝒉𝒚𝒑𝒆𝒓𝒃𝒐𝒍𝒆 𝒕𝒓è𝒔 é𝒕𝒊𝒓é𝒆 𝒆𝒏 𝒂𝒃𝒔𝒄𝒊𝒔𝒔𝒆 (𝒍𝒐𝒏𝒈𝒖𝒆 𝒒𝒖𝒆𝒖𝒆 𝒐𝒖 𝒉𝒆𝒂𝒗𝒚𝒕𝒂𝒊𝒍𝒆𝒅)𝒆𝒕 𝒐𝒓𝒅𝒐𝒏𝒏é𝒆
(𝒉𝒂𝒖𝒕𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 à 𝒍𝒂 𝒃𝒂𝒔𝒆). ..

𝓑– 𝟖 • 𝑳𝒐𝒊 𝒅𝒆 𝑳𝒂𝒑𝒍𝒂𝒄𝒆 𝒅𝒆 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆 𝝀 > 𝟎, 𝓛( 𝝀) ∶

𝟏 − |𝒙|
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶ ∀𝒙 ∈ ℝ , 𝒇𝑿 (𝒙) = 𝒆 𝝀
𝟐𝝀
𝟏 𝒙
𝒆𝝀 , 𝒔𝒊 𝒙 < 𝟎
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 ∶ 𝑭𝑿 (𝒙) = { 𝟐
𝟏 𝒙
𝟏 − 𝒆− 𝝀 , 𝒔𝒊 𝒙 ≥ 𝟎
𝟐
 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) = 𝟎  𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝑿) = 𝟐𝝀𝟐
𝟏
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 ∶ 𝑴𝑿 (𝒕) =
𝟏 − 𝝀𝟐 𝒕𝟐
𝓑– 𝟗 • 𝑳𝒐𝒊 𝒅𝒆 𝑮𝒖𝒎𝒃𝒆𝒍 𝒅𝒆 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆𝒔 𝝁 ∈ ℝ 𝒆𝒕 𝜷 > 0 , 𝔊(𝝁 , 𝜷):
𝑳𝒂 𝒍𝒐𝒊 𝒅𝒆 𝑮𝒖𝒎𝒃𝒆𝒍 𝒆𝒔𝒕 𝒖𝒕𝒊𝒍𝒊𝒔é𝒆 𝒑𝒐𝒖𝒓 𝒎𝒐𝒅é𝒍𝒊𝒔𝒆𝒓 𝒍𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 𝒆𝒙𝒕𝒓ê𝒎𝒆𝒔.
𝑷𝒂𝒓 𝒆𝒙𝒆𝒎𝒑𝒍𝒆, 𝒑𝒐𝒖𝒓 𝒅é𝒇𝒊𝒏𝒊𝒓 𝒅𝒆 𝒎𝒂𝒏𝒊è𝒓𝒆 𝒂𝒅é𝒒𝒖𝒂𝒕𝒆 𝒍𝒂 𝒑𝒖𝒊𝒔𝒔𝒂𝒏𝒄𝒆 𝒅′ 𝒖𝒏 𝒔𝒆𝒓𝒗𝒆𝒖𝒓, 𝒐𝒏
𝒔′ 𝒊𝒏𝒕é𝒓𝒆𝒔𝒔𝒆 𝒂𝒖 𝒏𝒐𝒎𝒃𝒓𝒆 𝒎𝒂𝒙𝒊𝒎𝒂𝒍 𝒅′ 𝒂𝒄𝒄è𝒔 𝒔𝒊𝒎𝒖𝒍𝒕𝒂𝒏é𝒔 à 𝒖𝒏 𝒔𝒊𝒕𝒆 𝒘𝒆𝒃 𝒅𝒂𝒏𝒔 𝒖𝒏𝒆 𝒋𝒐𝒖𝒓𝒏é𝒆,
𝒐𝒃𝒔𝒆𝒓𝒗é 𝒔𝒖𝒓 𝒑𝒍𝒖𝒔𝒊𝒆𝒖𝒓𝒔 𝒋𝒐𝒖𝒓𝒔.
𝒛 𝒙−𝝁
− −( )
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶ ∀𝒙 ∈ ℝ , 𝒇𝑿 (𝒙) = 𝒛𝒆 𝜷 𝒐ù 𝒛 = 𝒆 𝜷

𝒙−𝝁
−( )
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 ∶ ∀𝒙 ∈ ℝ , 𝑭𝑿 (𝒙) = 𝐞𝐱𝐩 (−𝒆 𝜷 )

 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) = 𝝁 + 𝜷𝜸 𝒐ù 𝜸 𝒆𝒔𝒕 𝒍𝒂 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆 𝒅′ 𝑬𝒖𝒍𝒆𝒓

+∞
𝟏 𝟏
𝜸 = ∑ [ − 𝐥𝐧 (𝟏 + )] ≅ 𝟎, 𝟓𝟕𝟕 𝟐𝟏𝟓 𝟔𝟔𝟒 𝟗
𝒌 𝒌
𝒌=𝟏

𝝅𝟐 𝟐
 𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝑿) = 𝜷
𝟔

295 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝓑– 𝟏𝟎 • 𝑳𝒐𝒊 𝑮𝒂𝒎𝒎𝒂 𝚪(𝜶, 𝜷) , 𝜶 > 𝟎 , 𝜷 > 𝟎 ∶
𝑳𝒂 𝒍𝒐𝒊 𝜞(𝜶, 𝜷)𝒅é𝒄𝒓𝒊𝒕 𝒍′ 𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆 𝒅𝒆 𝒕𝒆𝒎𝒑𝒔 𝒆𝒏𝒕𝒓𝒆 𝒍𝒆 𝒑𝒓𝒆𝒎𝒊𝒆𝒓 𝒆𝒕 𝒍𝒆 𝒅𝒆𝒓𝒏𝒊𝒆𝒓 é𝒗è𝒏𝒆𝒎𝒆𝒏𝒕 𝒅′ 𝒖𝒏𝒆
𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 (𝜶 + 𝟏) é𝒗è𝒏𝒆𝒎𝒆𝒏𝒕𝒔 𝒔𝒖𝒄𝒄𝒆𝒔𝒔𝒊𝒇𝒔.
𝑶𝒏 𝒑𝒆𝒖𝒕 𝒍𝒂 𝒄𝒐𝒏𝒔𝒊𝒅é𝒓𝒆𝒓 𝒄𝒐𝒎𝒎𝒆 𝒖𝒏𝒆 𝒈é𝒏é𝒓𝒂𝒍𝒊𝒔𝒂𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒆𝒙𝒑𝒐𝒏𝒆𝒏𝒕𝒊𝒆𝒍𝒍𝒆
𝒙
−
𝒙𝜶−𝟏 𝒆 𝜷
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶ 𝒇𝑿 (𝒙, 𝜶, 𝜷) = { 𝜞(𝜶)𝜷𝜶 , 𝒔𝒊 𝒙 ∈ [𝟎, +∞[
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
𝟏 𝝀𝜶 𝒆− 𝝀𝒙 𝜶−𝟏
𝑷𝒐𝒖𝒓 𝝀 = , ∀𝒙 ∈ [𝟎, +∞[ , 𝒇𝑿 (𝒙, 𝜶, 𝝀) = ( )𝒙
𝜷 𝜞(𝜶)

 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅𝒆 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝑮𝒂𝒎𝒎𝒂 𝒅′ 𝑬𝒖𝒍𝒆𝒓 𝜞(𝜶) ∶

+∞
𝓪 • ∀𝜶 > 𝟎 , 𝜞(𝜶) = ∫ 𝒙𝜶−𝟏 𝒆−𝒙 𝒅𝒙 𝓫 • ∀𝜶 > 𝟎 , 𝜞(𝜶 + 𝟏) = 𝜶𝜞(𝜶)
𝟎

𝜞(𝜶 + 𝒏)
𝓬 • ∀𝜶 > 𝟎 , ∀𝒏 ∈ ℕ∗ , 𝜞(𝜶) =
(𝜶 + 𝒏 − 𝟏) … (𝜶 + 𝟏)𝜶
𝓭 • ∀𝒏 ∈ ℕ , 𝜞(𝒏 + 𝟏) = 𝒏! , 𝒆𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 𝜞(𝟏) = 𝜞(𝟐) = 𝟏

𝟏 (𝟐𝒏)! √𝝅 𝟏 𝟑 √𝝅
𝓮 • ∀𝒏 ∈ ℕ , 𝜞 ( + 𝒏) = 𝟐𝒏 , 𝒆𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 𝜞 ( ) = √𝝅 , 𝜞 ( ) =
𝟐 𝟐 (𝒏!) 𝟐 𝟐 𝟐
 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) = 𝜶𝜷  𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝑿) = 𝜶𝜷𝟐
𝟐
 𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒔𝒚𝒎é𝒕𝒓𝒊𝒆 ∶ 𝜸𝟏 =
√𝜶
𝟔
 𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒑𝒍𝒂𝒕𝒊𝒔𝒔𝒆𝒎𝒆𝒏𝒕 (𝒌𝒖𝒓𝒕𝒐𝒔𝒊𝒔) ∶ 𝜸𝟐 =
𝜶
𝟏
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 ∶ 𝑴𝑿 (𝒕) = (𝟏 − 𝜷𝒕)−𝜶 , 𝒑𝒐𝒖𝒓 𝒕 <
𝜷
 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝑮𝒂𝒎𝒎𝒂 𝚪(𝜶, 𝜷) ∶
𝓪 • 𝑺𝒊 𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 𝒏 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒕𝒆𝒍𝒔 𝒒𝒖𝒆 ∀𝒊 = 𝟏, 𝟐, … , 𝒏 ; 𝑿𝒊 ↝ 𝜞(𝜶𝒊 , 𝜷) 𝒂𝒍𝒐𝒓𝒔 ,
𝒏 𝒏

∑ 𝑿𝒊 ↝ 𝜞 (∑ 𝜶𝒊 , 𝜷)
𝒊=𝟏 𝒊=𝟏

𝓫 • 𝑺𝒊 𝑿 ↝ 𝜞(𝜶, 𝜷), 𝒂𝒍𝒐𝒓𝒔 ∀𝒂 > 𝟎 , 𝒂𝑿 ↝ 𝜞(𝜶, 𝒂𝜷)

𝟏 𝟏
𝓬 • 𝑿 ↝ 𝜞 (𝟏, ) ⇔ 𝑿 ↝ 𝓔(𝝀) 𝒂𝒖𝒕𝒓𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒕 ∶ 𝜞 (𝟏, ) ≡ 𝓔(𝝀)
𝝀 𝝀
𝒏 𝒏
𝓭 • 𝑷𝒐𝒖𝒓 𝒏 ∈ ℕ∗ , 𝑿 ↝ 𝜞 ( , 𝟐) ⇔ 𝑿 ↝ 𝝌𝟐 (𝒏) 𝒂𝒖𝒕𝒓𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒕 ∶ 𝜞 ( , 𝟐) ≡ 𝝌𝟐 (𝒏) ,
𝟐 𝟐
𝝌𝟐 (𝒏) é𝒕𝒂𝒏𝒕 𝒍𝒂 𝒍𝒐𝒊 𝒌𝒉𝒊‑𝒅𝒆𝒖𝒙 à 𝒏 𝒅𝒆𝒈𝒓é𝒔 𝒅𝒆 𝒍𝒊𝒃𝒆𝒓𝒕é

296 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝓮 • 𝑺𝒊 𝜶 ≫ , 𝒂𝒍𝒐𝒓𝒔 𝜞(𝜶, 𝜷) ≈ 𝓝(𝒎, 𝝈𝟐 ) , 𝒂𝒗𝒆𝒄 𝒎 = (𝜶 − 𝟏)𝜷 𝒆𝒕 𝝈𝟐 = (𝜶 − 𝟏)𝜷𝟐

𝓑– 𝟏𝟏 • 𝑳𝒐𝒊 𝒃ê𝒕𝒂 𝕭(𝜶, 𝜷), 𝜶 > 𝟎 𝒆𝒕 𝜷 > 𝟎 ∶

𝒙𝜶−𝟏 (𝟏 − 𝒙)𝜷−𝟏
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶ 𝒇𝑿 (𝒙, 𝜶, 𝜷) = { , 𝒔𝒊 𝒙 ∈ [𝟎, 𝟏]
𝔹(𝜶, 𝜷)
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅𝒆 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝑩ê𝒕𝒂 𝔹(𝜶, 𝜷) ∶
𝟏
𝜞(𝜶)𝜞(𝜷)
𝓪 • ∀𝜶 > 𝟎 , 𝜷 > 𝟎 , 𝔹(𝜶, 𝜷) = ∫ 𝒕𝜶−𝟏 (𝟏 − 𝒕)𝜷−𝟏 𝒅𝒕 𝓫 • 𝔹(𝜶, 𝜷) = 𝔹(𝜷, 𝜶) =
𝟎 𝜞(𝜶 + 𝜷)
𝜷 𝟏
𝓬 • 𝔹(𝜶, 𝜷 + 𝟏) = ( ) 𝔹(𝜶, 𝜷) 𝓭 • 𝔹(𝜶, 𝜶) = 𝟐𝟏−𝟐𝜶 𝔹 ( , 𝜶)
𝜶+𝜷 𝟐
𝜶
 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) =
𝜶+𝜷
𝜶𝜷
 𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝑿) =
(𝜶 + 𝜷)𝟐 (𝜶
+ 𝜷 + 𝟏)
 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒃ê𝒕𝒂 𝕭(𝜶, 𝜷) ∶
𝓪 • 𝑿 ↝ 𝕭(𝟏, 𝟏) ⇔ 𝑿 ↝ 𝓤[𝟎,𝟏] 𝒂𝒖𝒕𝒓𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒕 ∶ 𝕭(𝟏, 𝟏) ≡ 𝓤[𝟎,𝟏]

𝑿 ↝ 𝕭(𝜶, 𝜽) 𝑿
𝓫 • 𝑺𝒊 {𝒀 ↝ 𝕭(𝜷, 𝜽) , 𝒂𝒍𝒐𝒓𝒔 ↝ 𝕭(𝜶, 𝜷)
𝑿+𝒀
𝑿 𝒆𝒕 𝒀 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒅𝒂𝒏𝒕𝒆𝒔
𝟏
𝓬 • 𝑺𝒊 𝑿 ↝ 𝓤[𝟎,𝟏] , 𝒂𝒍𝒐𝒓𝒔 𝑿𝟐 ↝ 𝕭 ( , 𝟏)
𝟐
𝓑– 𝟏𝟐 • 𝑳𝒐𝒊 𝒅𝒆 𝑷𝒆𝒂𝒓𝒔𝒐𝒏 𝒐𝒖 𝒌𝒉𝒊‑𝒅𝒆𝒖𝒙 à 𝒏 𝒅𝒆𝒈𝒓é𝒔 𝒅𝒆 𝒍𝒊𝒃𝒆𝒓𝒕é 𝝌𝟐 (𝒌) , 𝒌 ∈ ℕ∗ ∶
𝒌 𝒙
𝒙(𝟐−𝟏) 𝒆− 𝟐
, 𝒔𝒊 𝒙 ∈ [𝟎, +∞[
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶ 𝒇𝑿 (𝒙) = 𝒌 𝒌𝟐
𝜞 (𝟐) 𝟐
{ 𝟎 , 𝒔𝒊 𝒏𝒐𝒏
 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) = 𝒌  𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝑿) = 𝟐𝒌

 𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒔𝒚𝒎é𝒕𝒓𝒊𝒆 ∶ 𝜸𝟏 = √𝟖⁄𝒌

 𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒑𝒍𝒂𝒕𝒊𝒔𝒔𝒆𝒎𝒆𝒏𝒕 (𝒌𝒖𝒓𝒕𝒐𝒔𝒊𝒔) ∶ 𝜸𝟐 = 𝟏𝟐⁄𝒌
𝒌
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 ∶ 𝑴𝑿 (𝒕) = (𝟏 − 𝟐𝒕)−𝟐 , 𝒑𝒐𝒖𝒓 𝟐𝒕 < 1
 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒌𝒉𝒊‑𝒅𝒆𝒖𝒙 à 𝒏 𝒅𝒆𝒈𝒓é𝒔 𝒅𝒆 𝒍𝒊𝒃𝒆𝒓𝒕é 𝝌𝟐 (𝒌) ∶
𝒏

𝓪 • 𝑺𝒐𝒊𝒕 (𝒁𝟏 , 𝒁𝟐 , … , 𝒁𝒏 ) 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒊. 𝒊. 𝒅 𝒅𝒆 𝒗. 𝒂. 𝒅𝒆 𝓝(𝟎, 𝟏) , 𝒂𝒍𝒐𝒓𝒔 𝑿 = ∑ 𝒁𝟐𝒊 ↝ 𝝌𝟐 (𝒏)

𝒊=𝟏

297 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒌 𝒌
𝓫 • 𝑷𝒐𝒖𝒓 𝒌 ∈ ℕ∗ , 𝑿 ↝ 𝜞 ( , 𝟐) ⇔ 𝑿 ↝ 𝝌𝟐 (𝒌) 𝒂𝒖𝒕𝒓𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒕 ∶ 𝜞 ( , 𝟐) ≡ 𝝌𝟐 (𝒌)
𝟐 𝟐
𝟏 𝒌
𝓬 • 𝑺𝒊 𝑿 ↝ 𝝌𝟐 (𝒌), 𝒂𝒍𝒐𝒓𝒔 𝑿 ↝ 𝜞 ( , 𝟏)
𝟐 𝟐
𝓭 • 𝑺𝒊 𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 𝒏 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒕𝒆𝒍𝒔 𝒒𝒖𝒆 ∀𝒊 = 𝟏, 𝟐, … , 𝒏 ; 𝑿𝒊 ↝ 𝝌𝟐 (𝒌𝒊 ) 𝒂𝒍𝒐𝒓𝒔
𝒏 𝒏

∑ 𝑿𝒊 ↝ 𝝌𝟐 (∑ 𝒌𝒊 )
𝒊=𝟏 𝒊=𝟏

𝒌
𝓮 • 𝑺𝒊 𝑿 ↝ 𝝌𝟐 (𝒌), 𝒂𝒍𝒐𝒓𝒔 ∀𝒂 > 𝟎 , 𝒂𝑿 ↝ 𝜞 ( , 𝟐𝒂)
𝟐
𝟏 −𝟏 𝟐
𝓯 • 𝑺𝒊 𝑿 ↝ 𝝌𝟐 (𝒌), 𝒌 > 𝟓𝟎 𝒂𝒍𝒐𝒓𝒔 √𝟐𝑿 − √𝟐𝒌 − 𝟏 ↝ 𝓝(𝟎, 𝟏) 𝒆𝒕 𝝌𝟐𝜶 (𝒌) = [𝚽 (𝛂) + √𝟐𝒌 − 𝟏]
𝟐
𝓑– 𝟏𝟑 • 𝑳𝒐𝒊 𝒅𝒆 𝑺𝒕𝒖𝒅𝒆𝒏𝒕 à 𝒌 𝒅𝒆𝒈𝒓é𝒔 𝒅𝒆 𝒍𝒊𝒃𝒆𝒓𝒕é 𝓣(𝒌) , 𝒌 ∈ ℕ∗ ∶
 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶
𝒌+𝟏 −(
𝒌+𝟏
) −(
𝒌+𝟏
)
𝟏 𝜞( 𝟐 ) 𝒙𝟐 𝟐 𝟏 𝒙𝟐 𝟐
∀𝒙 ∈ ℝ , 𝒇𝑿 (𝒙) = (𝟏 + ) = (𝟏 + )
√𝒌𝝅 𝜞 (𝒌) 𝒌 𝟏 𝒌
√𝒌𝔹 (𝟐 , 𝟐)
𝒌
𝟐
 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) = 𝟎 , 𝒔𝒊 𝒌 > 𝟏
𝒌
 𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆: 𝑽𝒂𝒓(𝑿) = , 𝒔𝒊 𝒌 > 𝟐  𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒔𝒚𝒎é𝒕𝒓𝒊𝒆: 𝜸𝟏 = 𝟎 , 𝒔𝒊 𝒌 > 𝟑
𝒌−𝟐
𝟔
 𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅′ 𝒂𝒑𝒍𝒂𝒕𝒊𝒔𝒔𝒆𝒎𝒆𝒏𝒕 (𝒌𝒖𝒓𝒕𝒐𝒔𝒊𝒔) ∶ 𝜸𝟐 = , 𝒔𝒊 𝒌 > 𝟒
𝒌−𝟒
 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝑺𝒕𝒖𝒅𝒆𝒏𝒕 à 𝒌 𝒅𝒆𝒈𝒓é𝒔 𝒅𝒆 𝒍𝒊𝒃𝒆𝒓𝒕é 𝓣(𝒌), 𝒌 ∈ ℕ∗ ∶
𝒁 ↝ 𝓝(𝟎, 𝟏)
𝒁
𝓪 • 𝑺𝒊 { 𝑿 ↝ 𝝌𝟐 (𝒌) , 𝒂𝒍𝒐𝒓𝒔 𝑻 = ↝ 𝓣(𝒌)
𝒁 𝒆𝒕 𝑿 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 √ 𝑿
𝒌
𝓫 • 𝑷𝒐𝒖𝒓 𝒌 = 𝟏, 𝑿 ↝ 𝓣(𝟏) ⇔ 𝑿 ↝ 𝓒(𝟎 , 𝟏) 𝒂𝒖𝒕𝒓𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒕 ∶
𝓣(𝟏) ≡ 𝓒(𝟎 , 𝟏) , 𝒐ù 𝓒(𝟎 , 𝟏) 𝒆𝒔𝒕 𝒍𝒂
𝒍𝒐𝒊 𝒅𝒆 𝑪𝒂𝒖𝒄𝒉𝒚 𝒅𝒆 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆𝒔 𝟎 𝒆𝒕 𝟏
𝓛
𝓬 • 𝓣(𝒌) → 𝓝(𝟎, 𝟏) , 𝒆𝒏 𝒑𝒓𝒂𝒕𝒊𝒒𝒖𝒆 𝒑𝒐𝒖𝒓 𝒏 > 𝟑𝟎

𝓑– 𝟏𝟒 • 𝑳𝒐𝒊 𝒅𝒆 𝑭𝒊𝒔𝒉𝒆𝒓‑𝑺𝒏𝒆𝒅𝒆𝒄𝒐𝒓 à 𝒎 𝒆𝒕 𝒏 𝒅𝒆𝒈𝒓é𝒔 𝒅𝒆 𝒍𝒊𝒃𝒆𝒓𝒕é 𝓕(𝒎, 𝒏) , 𝒎, 𝒏 ∈ ℕ∗ ∶

 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶
𝒎+𝒏 𝒎 𝒏 𝒎 𝒎 𝒏
𝜞( ) 𝒎 𝟐 𝒏𝟐 𝒙( 𝟐 −𝟏) 𝒎 𝟐 𝒏𝟐 𝒎
[ 𝟐 ][ 𝒙( 𝟐 −𝟏)
𝒇𝑿 (𝒙) = 𝒎 𝒏 𝒎+𝒏 ] = [ 𝒎 𝒏 ]
𝜞 ( 𝟐 ) 𝜞 (𝟐 ) (𝒎𝒙 + 𝒏)( 𝟐 ) 𝔹 ( 𝟐 , 𝟐) [ (
𝒎+𝒏 ]
)
, 𝒔𝒊 𝒙 ≥ 𝟎
(𝒎𝒙 + 𝒏) 𝟐
{𝟎 , 𝒔𝒊 𝒏𝒐𝒏

298 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒏
 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 ∶ 𝑬(𝑿) = , 𝒔𝒊 𝒏 > 𝟐
𝒏−𝟐
𝟐𝒏𝟐 (𝒎 + 𝒏 − 𝟐)
 𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽𝒂𝒓(𝑿) = , 𝒔𝒊 𝒏 > 𝟒
𝒎(𝒏 − 𝟐)𝟐 (𝒏 − 𝟒)
 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝑭𝒊𝒔𝒉𝒆𝒓‑𝑺𝒏𝒆𝒅𝒆𝒄𝒐𝒓 à 𝒎 𝒆𝒕 𝒏 𝒅𝒆𝒈𝒓é𝒔 𝒅𝒆 𝒍𝒊𝒃𝒆𝒓𝒕é 𝓕(𝒎, 𝒏)
𝑿𝟏 ↝ 𝝌𝟐 (𝒌𝟏 )
𝑿𝟏 ⁄𝒌𝟏
𝓪 • 𝑺𝒊 {𝑿𝟐 ↝ 𝝌𝟐 (𝒌𝟐 ) , 𝒂𝒍𝒐𝒓𝒔 𝑭 = ↝ 𝓕(𝒌𝟏 , 𝒌𝟐 )
𝑿𝟐 ⁄𝒌𝟐
𝑿𝟏 𝒆𝒕 𝑿𝟐 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔
𝓫 • 𝑺𝒊 𝑿 ↝ 𝓣(𝒌), 𝒂𝒍𝒐𝒓𝒔 𝑿𝟐 ↝ 𝓕(𝟏, 𝒌)
𝟏 𝟏
𝓬 • 𝑺𝒊 𝑿 ↝ 𝓕(𝒎, 𝒏) , 𝒂𝒍𝒐𝒓𝒔 ↝ 𝓕(𝒏, 𝒎) 𝒂𝒊𝒏𝒔𝒊 𝓕𝟏−𝜶 (𝒎, 𝒏) =
𝑿 𝓕𝜶 (𝒏, 𝒎)
𝓭 • 𝑺𝒊 𝑿 ↝ 𝓕(𝒎, 𝒏) , 𝒂𝒍𝒐𝒓𝒔 𝒀 = 𝐥𝐢𝐦 𝒏𝑿 ↝ 𝝌𝟐 (𝒎)
𝒏→+∞

𝓮 • 𝑺𝒊 𝑿 ↝ 𝓝(𝟎, 𝟏) , 𝒂𝒍𝒐𝒓𝒔 𝑿𝟐 ↝ 𝓕(𝟏, +∞)

𝓒– 𝟏 • 𝑰𝒏é𝒈𝒂𝒍𝒊𝒕é 𝒅𝒆 𝑴𝒂𝒓𝒌𝒐𝒗 ∶
𝓪 • 𝑺𝒊 𝑿 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒑𝒐𝒔𝒊𝒕𝒊𝒗𝒆 𝒂𝒚𝒂𝒏𝒕 𝒖𝒏𝒆 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 , 𝒐𝒏 𝒂 ∶
𝑬(𝑿)
∀𝒕 > 𝟎 , 𝑷(𝑿 ≥ 𝒕) ≤
𝒕
𝓫 • 𝑺𝒊 𝑿𝒆𝒔𝒕 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒂𝒚𝒂𝒏𝒕 𝒖𝒏 𝒎𝒐𝒎𝒆𝒏𝒕 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝒓 ∶
𝑬(|𝑿|𝒓 )
∀𝒕 > 𝟎 , 𝑷(|𝑿| ≥ 𝒕) ≤
𝒕𝒓
𝓒– 𝟐 • 𝑰𝒏é𝒈𝒂𝒍𝒊𝒕é 𝒅𝒆 𝑩𝒊𝒆𝒏𝒂𝒚𝒎é‑𝑻𝒄𝒉𝒆𝒃𝒚𝒄𝒉𝒆𝒗 ∶
𝑽𝒂𝒓(𝑿)
𝑺𝒐𝒊𝒕 𝑿 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒓é𝒆𝒍𝒍𝒆 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒇𝒊𝒏𝒊𝒆. ∀ 𝝐 > 𝟎 , 𝑷(|𝑿 − 𝑬(𝑿)| ≥ 𝝐) ≤
𝝐𝟐
𝓒– 𝟑 • 𝑪𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒄𝒆 𝒑𝒓𝒆𝒔𝒒𝒖𝒆 𝒔û𝒓𝒆 ∶
𝓪 • 𝑫é𝒇𝒊𝒏𝒊𝒕𝒊𝒐𝒏 𝟏 ∶ 𝑺𝒐𝒊𝒕 (𝑿𝒏 )𝒏≥𝟏 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒆𝒕 𝑿 𝒖𝒏𝒆 𝒗. 𝒂 𝒅é𝒇𝒊𝒏𝒊𝒆𝒔 𝒔𝒖𝒓 𝒍𝒆 𝒎ê𝒎𝒆
𝒆𝒔𝒑𝒂𝒄𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒔é (𝛀, 𝓕, 𝓟) . 𝑶𝒏 𝒅𝒊𝒕 𝒒𝒖𝒆 𝑿𝒏 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒑𝒓𝒆𝒔𝒒𝒖𝒆 𝒔û𝒓𝒆𝒎𝒆𝒏𝒕 𝒗𝒆𝒓𝒔 𝑿 ,
𝒔𝒊 𝒍′ 𝒆𝒏𝒔𝒆𝒎𝒃𝒍𝒆 𝒅𝒆𝒔 𝝎 𝒕𝒆𝒍𝒔 𝒒𝒖𝒆 𝑿𝒏 (𝝎) 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒗𝒆𝒓𝒔 𝑿(𝝎) 𝒂 𝒑𝒐𝒖𝒓 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝟏
𝒑.𝒔
𝑶𝒏 𝒏𝒐𝒕𝒆 𝑿𝒏 → 𝑿
𝒏→+∞
𝒑.𝒔
𝓫 • 𝑫é𝒇𝒊𝒏𝒊𝒕𝒊𝒐𝒏 𝟐 ∶ 𝑳𝒂 𝒔𝒖𝒊𝒕𝒆 (𝑿𝒏 ) 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒑𝒓𝒆𝒔𝒒𝒖𝒆 𝒔û𝒓𝒆𝒎𝒆𝒏𝒕 𝒗𝒆𝒓𝒔 𝑿 , 𝒏𝒐𝒕é 𝑿𝒏 → 𝑿
𝒏→+∞

299 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒔𝒊 𝑷{𝝎: 𝑿𝒏 ↛ 𝑿} = 𝟎 (𝒆𝒏 𝒕𝒉é𝒐𝒓𝒊𝒆 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒐𝒊𝒕é ∶ ∀𝝐 > 𝟎 , 𝐥𝐢𝐦 𝑷 {𝐬𝐮𝐩|𝑿𝒌 − 𝑿| ≥ 𝝐} = 𝟎
𝒏→+∞ 𝒌≥𝒏

𝓒– 𝟒 • 𝑪𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒄𝒆 𝒆𝒏 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶
𝑺𝒐𝒊𝒕 (𝑿𝒏 )𝒏≥𝟏 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒆𝒕 𝑿 𝒖𝒏𝒆 𝒗. 𝒂 𝒅é𝒇𝒊𝒏𝒊𝒆𝒔 𝒔𝒖𝒓 𝒍𝒆 𝒎ê𝒎𝒆 𝒆𝒔𝒑𝒂𝒄𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒔é (𝛀, 𝓕, 𝓟).
𝑶𝒏 𝒅𝒊𝒕 𝒒𝒖𝒆 𝑿𝒏 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒗𝒆𝒓𝒔 𝑿
𝒔𝒊 ∶ ∀𝝐 > 𝟎 , 𝐥𝐢𝐦 𝑷(|𝑿𝒏 − 𝑿| ≥ 𝝐) = 𝟎 𝒐𝒖 𝒆𝒏𝒄𝒐𝒓𝒆 ∀𝝐 > 𝟎 , 𝐥𝐢𝐦 𝑷(|𝑿𝒏 − 𝑿| < 𝝐) = 𝟏
𝒏→+∞ 𝒏→+∞
𝒑
𝑶𝒏 𝒏𝒐𝒕𝒆 𝐩𝐥𝐢𝐦 (𝑿𝒏 ) = 𝑿 𝒐𝒖 𝒆𝒏𝒄𝒐𝒓𝒆 𝑿𝒏 → 𝑿
𝒏→+∞ 𝒏→+∞

𝒑 𝒑
𝓪 • 𝑻𝒉é𝒐𝒓è𝒎𝒆 𝒅𝒆 𝑺𝒍𝒖𝒕𝒔𝒌𝒚 ∶ 𝑿𝒏 → 𝑿 ⇒ 𝒈(𝑿𝒏 ) → 𝒈(𝑿) , 𝒐ù 𝒈 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏
𝒏→+∞ 𝒏→+∞

𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒅𝒆 ℝ 𝒅𝒂𝒏𝒔 ℝ
𝒑.𝒔 𝒑
𝓫 • 𝑿𝒏 → 𝑿 ⇒ 𝑿𝒏 → 𝑿
𝒏→+∞ 𝒏→+∞

𝓬 • 𝑪𝒐𝒏𝒔𝒊𝒅é𝒓𝒐𝒏𝒔 𝒍𝒆𝒔 𝒔𝒖𝒊𝒕𝒆𝒔 (𝑿𝒏 )𝒏≥𝟏 𝒆𝒕 (𝒀𝒏 )𝒏≥𝟏 𝒅𝒆 𝒗. 𝒂. 𝑺𝒊 𝒐𝒏 𝒂 𝒍𝒆𝒔 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒄𝒆𝒔

𝒑
𝑿𝒏 → 𝑿 𝒑 𝒑
𝒏→+∞
{ 𝒑 , 𝒂𝒍𝒐𝒓𝒔 , 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕 𝝀 ∈ ℝ , 𝑿𝒏 + 𝝀𝒀𝒏 → 𝑿 + 𝝀𝒀 𝒆𝒕 𝑿𝒏 𝒀𝒏 → 𝑿𝒀
𝒏→+∞ 𝒏→+∞
𝒀𝒏 → 𝒀
𝒏→+∞

𝓒– 𝟒 • 𝑪𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒄𝒆 𝒆𝒏 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒒𝒖𝒂𝒅𝒓𝒂𝒕𝒊𝒒𝒖𝒆 ∶

𝑶𝒏 𝒅𝒊𝒕 𝒒𝒖′ 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. (𝑿𝒏 )𝒏≥𝟏 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒒𝒖𝒂𝒅𝒓𝒂𝒕𝒊𝒒𝒖𝒆 𝒗𝒆𝒓𝒔 𝒍𝒂 𝒗. 𝒂. 𝑿
𝒎𝒒
𝒆𝒕 𝒐𝒏 𝒏𝒐𝒕𝒆 ∶ 𝑿𝒏 → 𝑿 , 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 𝐥𝐢𝐦 𝑬(|𝑿𝒏 − 𝑿|𝟐 ) = 𝟎
𝒏→+∞ 𝒏→+∞

𝒎𝒒 𝐥𝐢𝐦 𝑬(𝑿𝒏 ) = 𝑿
𝑺𝒊 𝒍𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒅′ 𝒐𝒓𝒅𝒓𝒆𝟏 𝒆𝒕 𝟐 𝒆𝒙𝒊𝒔𝒕𝒆𝒏𝒕 , 𝒂𝒍𝒐𝒓𝒔 ∶ (𝑿𝒏 → 𝑿) ⇔ {𝒏→+∞
𝒏→+∞ 𝐥𝐢𝐦 𝑽𝒂𝒓(𝑿𝒏 ) = 𝟎
𝒏→+∞
𝒎𝒒 𝒑
𝓪 • 𝑿𝒏 → 𝑿 ⇒ 𝑿𝒏 → 𝑿
𝒏→+∞ 𝒏→+∞
𝒑.𝒔 𝒑
𝓫 • 𝑿𝒏 → 𝑿 ⇒ (𝑬(𝑿𝒏 ))𝒏≥𝟏 → (𝑿𝒏 )𝒏≥𝟏
𝒏→+∞ 𝒏→+∞

𝓒– 𝟓 • 𝑪𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒄𝒆 𝒆𝒏 𝒍𝒐𝒊 ∶
𝑺𝒐𝒊𝒕 (𝑿𝒏 )𝒏≥𝟏 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒆𝒕 𝑿 𝒖𝒏𝒆 𝒗. 𝒂 𝒅é𝒇𝒊𝒏𝒊𝒆𝒔 𝒔𝒖𝒓 𝒍𝒆 𝒎ê𝒎𝒆 𝒆𝒔𝒑𝒂𝒄𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒔é (𝛀, 𝓕, 𝓟).
𝓛
𝑶𝒏 𝒅𝒊𝒕 𝒒𝒖𝒆 𝑿𝒏 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒍𝒐𝒊 𝒗𝒆𝒓𝒔 𝑿 𝒆𝒕 𝒐𝒏 𝒏𝒐𝒕𝒆 𝑿𝒏 → 𝑿 , 𝒔𝒊 𝒆𝒏 𝒕𝒐𝒖𝒕 𝒑𝒐𝒊𝒏𝒕 𝒙 𝒐ù 𝑭𝑿 𝒆𝒔𝒕
𝒏→+∞

𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 , 𝒐𝒏 𝒂 𝐥𝐢𝐦 𝑭𝑿𝒏 (𝒙) = 𝑭𝑿 (𝒙)

𝒏→+∞

𝓪 • 𝑶𝒏 𝒄𝒐𝒏𝒔𝒕𝒂𝒕𝒆 𝒒𝒖𝒆 𝒍𝒂 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒄𝒆 𝒆𝒏 𝒍𝒐𝒊 𝒅𝒆 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 (𝑿𝒏 )𝒏≥𝟏 𝒗𝒆𝒓𝒔 𝑿 𝒆𝒔𝒕 é𝒒𝒖𝒊𝒗𝒂𝒍𝒆𝒏𝒕𝒆
à 𝐥𝐢𝐦 𝑷(𝒂 < 𝑿𝒏 ≤ 𝒃) = 𝑷(𝒂 < 𝑿 ≤ 𝒃) , 𝒐ù 𝒂 𝒆𝒕 𝒃 𝒔𝒐𝒏𝒕 𝒅𝒆𝒖𝒙 𝒑𝒐𝒊𝒏𝒕𝒔 𝒅𝒆 𝒄𝒐𝒏𝒕𝒖𝒏𝒖𝒊𝒕é 𝒅𝒆 𝑭𝑿
𝒏→+∞

𝓫 • 𝑺𝒊 (𝑿𝒏 )𝒏≥𝟏 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒅𝒊𝒔𝒄𝒓è𝒕𝒆𝒔 𝒆𝒕 𝒔𝒊 𝑿 𝒆𝒔𝒕 é𝒈𝒂𝒍𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒔𝒄𝒓è𝒕𝒆 𝒕𝒆𝒍𝒍𝒆 𝒒𝒖𝒆

300 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
∀𝒏 ∈ ℕ 𝑿𝒏 (𝛀) ⊂ 𝑿(𝛀) , 𝒂𝒍𝒐𝒓𝒔 (𝑿𝒏 )𝒏≥𝟏 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒍𝒐𝒊 𝒗𝒆𝒓𝒔 𝑿 , 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 ∶
∀𝒙 ∈ 𝑿(𝛀) , 𝐥𝐢𝐦 𝑷(𝑿𝒏 = 𝒙) = 𝑷(𝑿 = 𝒙)
𝒏→+∞
𝒎𝒒 𝒑 𝓛
𝓬 • 𝑿𝒏 → 𝑿 ⇒ 𝑿𝒏 → 𝑿 ⇒ 𝑿𝒏 → 𝑿
𝒏→+∞ 𝒏→+∞ 𝒏→+∞

𝓭 • 𝑺𝒊 (𝑿𝒏 )𝒏≥𝟏 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒂𝒏𝒕 𝒆𝒏 𝒍𝒐𝒊 𝒗𝒆𝒓𝒔 𝒖𝒏𝒆 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆 𝒂 𝒅𝒆 ℝ ,
𝓛 𝒑
𝒂𝒍𝒐𝒓𝒔 𝒆𝒍𝒍𝒆 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 é𝒈𝒂𝒍𝒆𝒎𝒆𝒏𝒕 𝒆𝒏 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒗𝒆𝒓𝒔 𝒍𝒂 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆 𝒂 ∶ 𝑿𝒏 → 𝒂 ⇒ 𝑿𝒏 → 𝒂
𝒏→+∞ 𝒏→+∞

𝓛
𝑿𝒏 → 𝑿
𝒏→+∞
𝓮 • 𝑺𝒖𝒑𝒑𝒐𝒔𝒐𝒏𝒔 𝒒𝒖𝒆 𝒍′ 𝒐𝒏 𝒂𝒊𝒕 𝒍𝒆𝒔 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒄𝒆𝒔 ∶ { 𝒑 , 𝒑𝒐𝒖𝒓 𝒂 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆 𝒓é𝒆𝒍𝒍𝒆.
𝒀𝒏 → 𝒂
𝒏→+∞

𝓛 𝓛 𝑿𝒏 𝓛 𝑿
𝑨𝒍𝒐𝒓𝒔, 𝒐𝒏 𝒂 ∶ ① 𝑿𝒏 + 𝒀𝒏 → 𝑿+𝒂 ② 𝑿𝒏 𝒀𝒏 → 𝒂𝑿 ③ → , 𝒔𝒊 𝒂 ≠ 𝟎
𝒏→+∞ 𝒏→+∞ 𝒀𝒏 𝒏→+∞ 𝒂
𝓯 • 𝑺𝒐𝒊𝒕 (𝑿𝑵 )𝑵≥𝟏 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒅𝒆 𝒍𝒐𝒊 𝒉𝒚𝒑𝒆𝒓𝒈é𝒐𝒎é𝒕𝒓𝒊𝒒𝒖𝒆 𝓗(𝑵, 𝒏, 𝒑). 𝑶𝒏 𝒏𝒐𝒕𝒆 𝑺 ,
𝒍′ 𝒆𝒏𝒔𝒆𝒎𝒃𝒍𝒆 𝒅𝒆𝒔 𝒏𝒐𝒎𝒃𝒓𝒆𝒔 𝒆𝒏𝒕𝒊𝒆𝒓𝒔 𝑵 𝒕𝒆𝒍𝒔 𝒒𝒖𝒆 𝑵𝒑 𝒔𝒐𝒊𝒕 𝒆𝒏𝒕𝒊𝒆𝒓 . 𝑨𝒍𝒐𝒓𝒔 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 (𝑿𝑵 )𝑵∈ℕ
𝓛
𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝑳𝒐𝒊 𝒗𝒆𝒓𝒔 𝒖𝒏𝒆 𝒗. 𝒂. 𝒃𝒊𝒏𝒐𝒎𝒊𝒂𝒍𝒆 𝓑(𝒏, 𝒑). 𝑶𝒏 𝒏𝒐𝒕𝒆 ∶ 𝑿𝑵 ↝ 𝓗(𝑵, 𝒏, 𝒑) ⇒ 𝑿𝑵 → 𝓑(𝒏, 𝒑)
𝑵→+∞

𝓰 • 𝑺𝒐𝒊𝒕 (𝑿𝒏 )𝒏∈ℕ 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒅𝒆 𝒍𝒐𝒊𝒔 𝒃𝒊𝒏𝒐𝒎𝒊𝒂𝒍𝒆𝒔 𝓑(𝒏, 𝒑𝒏 ) 𝒓𝒆𝒔𝒑𝒆𝒄𝒕𝒊𝒗𝒆𝒎𝒆𝒏𝒕.

𝑶𝒏 𝒔𝒖𝒑𝒑𝒐𝒔𝒆 𝒒𝒖𝒆 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 (𝒏𝒑𝒏 )𝒏∈ℕ 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒗𝒆𝒓𝒔 𝒖𝒏 𝒓é𝒆𝒍 𝒑𝒐𝒔𝒊𝒕𝒊𝒇 𝝀 . 𝑨𝒍𝒐𝒓𝒔 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 (𝑿𝒏 )𝒏∈ℕ
𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒍𝒐𝒊 𝒗𝒆𝒓𝒔 𝒖𝒏𝒆 𝒗. 𝒂. 𝒅𝒆 𝑷𝒐𝒊𝒔𝒔𝒐𝒏 𝓟(𝝀). 𝑶𝒏 𝒏𝒐𝒕𝒆 ∶
𝑿𝒏 ↝ 𝓑(𝒏, 𝒑𝒏 ) 𝓛
{ 𝐥𝐢𝐦 𝒏𝒑 = 𝝀 > 0 ⇒ 𝑿𝒏 → 𝓟(𝝀)
𝒏 𝒏→+∞
𝒏→+∞

𝓱 • 𝑻𝒉é𝒐𝒓è𝒎𝒆 𝒅𝒆 𝑴𝒐𝒊𝒗𝒓𝒆‑𝑳𝒂𝒑𝒍𝒂𝒄𝒆 ∶ 𝑺𝒐𝒊𝒕 (𝑿𝒏 )𝒏∈ℕ 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒅𝒆 𝒍𝒐𝒊𝒔 𝒃𝒊𝒏𝒐𝒎𝒊𝒂𝒍𝒆𝒔

𝓑(𝒏, 𝒑) 𝒓𝒆𝒔𝒑𝒆𝒄𝒕𝒊𝒗𝒆𝒎𝒆𝒏𝒕 . 𝑶𝒏 𝒄𝒐𝒏𝒔𝒊𝒅è𝒓𝒆 𝒍𝒂 𝒗. 𝒂. (𝑻𝒏 )𝒏∈ℕ∗ 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒑𝒂𝒓 ∶
𝑿𝒏 − 𝒏𝒑
∀ 𝒏 ∈ ℕ∗ , 𝑻𝒏 = .
√𝒏𝒑(𝟏 − 𝒑)
𝑨𝒍𝒐𝒓𝒔 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 (𝑿𝒏 )𝒏∈ℕ 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒍𝒐𝒊 𝒗𝒆𝒓𝒔 𝒖𝒏𝒆 𝒗. 𝒂. 𝒅𝒆 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝒄𝒆𝒏𝒕𝒓é𝒆 𝒓é𝒅𝒖𝒊𝒕𝒆 .
𝑿𝒏 − 𝒏𝒑 𝓛
𝑶𝒏 𝒏𝒐𝒕𝒆 ∶ 𝑿𝒏 ↝ 𝓑(𝒏, 𝒑) ⇒ → 𝓝(𝟎, 𝟏)
√𝒏𝒑(𝟏 − 𝒑) 𝒏→+∞
𝓲 • 𝑺𝒐𝒊𝒕 (𝑿𝒏 )𝒏∈ℕ 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒅𝒆 𝒍𝒐𝒊𝒔 𝒅𝒆 𝑷𝒐𝒊𝒔𝒔𝒐𝒏 𝓟(𝒏𝝀) 𝒓𝒆𝒔𝒑𝒆𝒄𝒕𝒊𝒗𝒆𝒎𝒆𝒏𝒕 .
𝑿𝒏 − 𝒏𝝀
𝑶𝒏 𝒄𝒐𝒏𝒔𝒊𝒅è𝒓𝒆 𝒍𝒂 𝒗. 𝒂. (𝑻𝒏 )𝒏∈ℕ∗ 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒑𝒂𝒓 ∶ ∀ 𝒏 ∈ ℕ∗ , 𝑻𝒏 = .
√𝒏𝝀
𝑨𝒍𝒐𝒓𝒔 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 (𝑿𝒏 )𝒏∈ℕ 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒍𝒐𝒊 𝒗𝒆𝒓𝒔 𝒖𝒏𝒆 𝒗. 𝒂. 𝒅𝒆 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝒄𝒆𝒏𝒕𝒓é𝒆 𝒓é𝒅𝒖𝒊𝒕𝒆 .
𝑿𝒏 − 𝒏𝝀 𝓛
𝑶𝒏 𝒏𝒐𝒕𝒆 ∶ 𝑿𝒏 ↝ 𝓟(𝒏𝝀) ⇒ → 𝓝(𝟎, 𝟏)
√𝒏𝝀 𝒏→+∞

301 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
(𝑿 )
𝓳 • 𝑺𝒐𝒊𝒕 𝒏 𝒏∈ℕ 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒆𝒕 (𝑴𝑿𝒏 ) 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆𝒔 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏𝒔 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆𝒔
𝓛
𝒂𝒔𝒔𝒐𝒄𝒊é𝒆𝒔. (𝑿𝒏 → 𝑿) ⇔ ( 𝐥𝐢𝐦 𝑴𝑿𝒏 (𝒕) = 𝑴𝑿 (𝒕))
𝒏→+∞ 𝒏→+∞

𝓒– 𝟔 • 𝑳𝒐𝒊 𝒇𝒂𝒊𝒃𝒍𝒆 𝒅𝒆𝒔 𝒈𝒓𝒂𝒏𝒅𝒔 𝒏𝒐𝒎𝒃𝒓𝒆𝒔 ∶

𝑺𝒐𝒊𝒕 (𝑿𝒏 )𝒏∈ℕ∗ 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒂𝒚𝒂𝒏𝒕 𝒎ê𝒎𝒆 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒆𝒕 𝒎ê𝒎𝒆 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆
𝒏
𝟏 𝒑
𝒇𝒊𝒏𝒊𝒆𝒔. 𝑨𝒍𝒐𝒓𝒔 ∶ ∑ 𝑿𝒊 → 𝑬(𝑿𝟏 )
𝒏 𝒏→+∞
𝒊=𝟏

𝓪 • 𝑺𝒐𝒊𝒕 (𝑿𝒏 )𝒏∈ℕ∗ 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒅𝒆 𝑩𝒆𝒓𝒏𝒐𝒖𝒍𝒍𝒊 𝒅𝒆 𝒎ê𝒎𝒆 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆 𝒑

𝒏
𝟏 𝒑
𝓑(𝟏, 𝒑). 𝑨𝒍𝒐𝒓𝒔 ∶ ∑ 𝑿𝒊 → 𝒑
𝒏 𝒏→+∞
𝒊=𝟏

𝓒– 𝟕 • 𝑳𝒐𝒊 𝒇𝒐𝒓𝒕𝒆 𝒅𝒆𝒔 𝒈𝒓𝒂𝒏𝒅𝒔 𝒏𝒐𝒎𝒃𝒓𝒆𝒔 ∶

𝑺𝒐𝒊𝒕 (𝑿𝒏 )𝒏∈ℕ∗ 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒆𝒕 𝒅𝒆 𝒎ê𝒎𝒆 𝒍𝒐𝒊 , 𝒂𝒗𝒆𝒄 𝒖𝒏 𝒎𝒐𝒎𝒆𝒏𝒕 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝟒
𝒏
𝟏 𝒑.𝒔
(𝒊. 𝒆. 𝑬(𝑿𝟒𝒊 ) < +∞) . 𝑨𝒍𝒐𝒓𝒔 ∶ ∑ 𝑿𝒊 → 𝑬(𝑿𝟏 )
𝒏 𝒏→+∞
𝒊=𝟏

𝓒– 𝟖 • 𝑻𝒉é𝒐𝒓è𝒎𝒆 𝒄𝒆𝒏𝒕𝒓𝒂𝒍 𝒍𝒊𝒎𝒊𝒕𝒆 ∶

𝑺𝒐𝒊𝒕 (𝑿𝒏 )𝒏∈ℕ∗ 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒆𝒕 𝒅𝒆 𝒎ê𝒎𝒆 𝒍𝒐𝒊 , 𝒅𝒆 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒎 𝒆𝒕 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆
𝝈𝟐 (𝒊. 𝒆. 𝒂𝒗𝒆𝒄 𝒖𝒏 𝒎𝒐𝒎𝒆𝒏𝒕 𝒅′ 𝒐𝒓𝒅𝒓𝒆 𝟐 𝒇𝒊𝒏𝒊).
𝒏
𝑿𝟏 + 𝑿𝟐 + ⋯ + 𝑿𝒏 𝟏
𝑵𝒐𝒕𝒐𝒏𝒔 𝑴𝒏 𝒍𝒆𝒔 𝒎𝒐𝒚𝒆𝒏𝒏𝒆𝒔 𝒂𝒓𝒊𝒕𝒉𝒎é𝒕𝒊𝒒𝒖𝒆𝒔 ∶ 𝑴𝒏 = = ∑ 𝑿𝒊
𝒏 𝒏
𝒊=𝟏

𝑴𝒏 − 𝒎 √𝒏(𝑴𝒏 − 𝒎)
𝒆𝒕 𝒁𝒏 𝒍𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒄𝒆𝒏𝒕𝒓é𝒆𝒔 𝒓é𝒅𝒖𝒊𝒕𝒆𝒔 𝒂𝒔𝒔𝒐𝒄𝒊é𝒆𝒔 ∶ 𝒁𝒏 = = .
𝝈⁄√𝒏 𝝈
𝒕𝟐
𝒃 − 𝟐
𝒆
𝑨𝒍𝒐𝒓𝒔 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕 𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆 [𝒂, 𝒃], 𝒐𝒏 𝒂 ∶ 𝐥𝐢𝐦 𝑷[𝒂 ≤ 𝒁𝒏 ≤ 𝒃] = ∫ 𝒅𝒕 .
𝒏→+∞ 𝒂 √𝟐𝝅
√𝒏(𝑴𝒏 − 𝒎)
𝑶𝒏 𝒅𝒊𝒕 𝒒𝒖𝒆 𝒍𝒂 𝒗. 𝒂. 𝒁𝒏 = 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒍𝒐𝒊 𝒗𝒆𝒓𝒔 𝒍𝒂 𝒍𝒐𝒊 𝑵𝒐𝒓𝒎𝒂𝒍𝒆
𝝈
√𝒏(𝑴𝒏 − 𝒎) 𝓛
𝒔𝒕𝒂𝒏𝒅𝒂𝒓𝒅 𝓝(𝟎, 𝟏) ∶ → 𝓝(𝟎, 𝟏)
𝝈 𝒏→+∞

302 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014

𝓒– 𝟏 • 𝑽𝒆𝒄𝒕𝒆𝒖𝒓𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 ∶
𝑶𝒏 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒄𝒐𝒖𝒑𝒍𝒆 (𝑿, 𝒀) 𝒅𝒆 𝒗. 𝒂. 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆𝒔 𝒅𝒆 ℝ𝟐 , 𝒔′𝒊𝒍 𝒆𝒙𝒊𝒔𝒕𝒆 𝒖𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒇: ℝ𝟐 ⟶ ℝ+
𝒕𝒆𝒍𝒍𝒆 𝒒𝒖𝒆 , 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒔 𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆𝒔 𝑰 𝒆𝒕 𝑱 𝒆𝒕 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒃𝒐𝒓𝒏é𝒆 𝒉 , 𝒐𝒏 𝒂𝒊𝒕 ∶

𝑷((𝑿, 𝒀) ∈ 𝑰 × 𝑱) = ∬ 𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚 𝒆𝒕 𝑬[𝒉(𝑿, 𝒀)] = ∬ 𝒉(𝒙, 𝒚)𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚
𝑰×𝑱 ℝ𝟐

𝓒– 𝟐 • 𝑫𝒆𝒏𝒔𝒊𝒕é 𝒄𝒐𝒏𝒋𝒐𝒊𝒏𝒕𝒆 ∶
𝒇𝑿,𝒀 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒄𝒐𝒏𝒋𝒐𝒊𝒏𝒕𝒆 𝒅𝒖 𝒄𝒐𝒖𝒑𝒍𝒆 (𝑿, 𝒀) 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 ∶
+∞ +∞
∫ ∫ 𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚 = 𝟏 𝒆𝒕 𝒇𝑿,𝒀 (𝒙, 𝒚) ≥ 𝟎 , ∀(𝒙, 𝒚) ∈ ℝ𝟐
−∞ −∞

𝓒– 𝟑 • 𝑫𝒆𝒏𝒔𝒊𝒕é𝒔 𝒎𝒂𝒓𝒈𝒊𝒏𝒂𝒍𝒆𝒔 ∶
𝑺𝒐𝒊𝒕 (𝑿, 𝒀) 𝒆𝒔𝒕 𝒖𝒏 𝒄𝒐𝒖𝒑𝒍𝒆 𝒆 𝒗. 𝒂. 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆𝒔 𝒅𝒆 ℝ𝟐 , 𝒔𝒆𝒔 𝒅𝒆𝒏𝒔𝒊𝒕é𝒔 𝒎𝒂𝒓𝒈𝒊𝒏𝒂𝒍𝒆𝒔 𝒇𝑿 𝒆𝒕 𝒇𝒀 𝒑𝒆𝒖𝒗𝒆𝒏𝒕
𝒔𝒆 𝒄𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒑𝒂𝒓 ∶

 𝒇𝑿 (𝒙) = ∫ 𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒚  𝒇𝒀 (𝒚) = ∫ 𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙

𝒚∈ℝ 𝒙∈ℝ

 𝑹𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒔 ∶ 𝑳𝒂 𝒄𝒐𝒏𝒏𝒂𝒊𝒔𝒔𝒂𝒏𝒄𝒆 𝒅𝒆 𝒍𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒄𝒐𝒏𝒋𝒐𝒊𝒏𝒕𝒆 𝒅𝒆 (𝑿, 𝒀) 𝒑𝒆𝒓𝒎𝒆𝒕 𝒅𝒆

𝒄𝒂𝒍𝒄𝒖𝒍𝒆𝒓𝒍𝒆𝒔 𝒅𝒆𝒏𝒔𝒊𝒕é𝒔 𝒎𝒂𝒓𝒈𝒊𝒏𝒂𝒍𝒆𝒔. 𝑰𝒍 𝒊𝒎𝒑𝒐𝒓𝒕𝒆 𝒅𝒆 𝒃𝒊𝒆𝒏 𝒄𝒐𝒎𝒑𝒓𝒆𝒏𝒅𝒓𝒆 𝒒𝒖𝒆 𝒍𝒂 𝒓é𝒄𝒊𝒑𝒓𝒐𝒒𝒖𝒆 𝒆𝒔𝒕
𝒇𝒂𝒖𝒔𝒔𝒆. 𝑰𝒍 𝒏’𝒆𝒔𝒕 𝒈é𝒏é𝒓𝒂𝒍𝒆𝒎𝒆𝒏𝒕 𝒑𝒂𝒔 𝒑𝒐𝒔𝒔𝒊𝒃𝒍𝒆 𝒅𝒆 𝒄𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒇𝑿,𝒀 𝒅𝒖 𝒄𝒐𝒖𝒑𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆
(𝑿, 𝒀) à 𝒑𝒂𝒓𝒕𝒊𝒓 𝒅𝒆 𝒍𝒂 𝒔𝒆𝒖𝒍𝒆 𝒄𝒐𝒏𝒏𝒂𝒊𝒔𝒔𝒂𝒏𝒄𝒆 𝒅𝒆 𝒔𝒆𝒔 𝒅𝒆𝒏𝒔𝒊𝒕é𝒔 𝒎𝒂𝒓𝒈𝒊𝒏𝒂𝒍𝒆𝒔 𝒇𝑿 𝒆𝒕 𝒇𝒀
 𝑨𝒕𝒕𝒆𝒏𝒕𝒊𝒐𝒏 ∶ 𝑫𝒆𝒖𝒙 𝒗𝒆𝒄𝒕𝒆𝒖𝒓𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒅𝒊𝒇𝒇é𝒓𝒆𝒏𝒕𝒔 𝒑𝒆𝒖𝒗𝒆𝒏𝒕 𝒂𝒗𝒐𝒊𝒓 𝒍𝒆𝒔 𝒎ê𝒎𝒆𝒔
𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏𝒔 𝒎𝒂𝒓𝒈𝒊𝒏𝒂𝒍𝒆𝒔.

𝓒– 𝟒 • 𝑰𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒄𝒆 𝒅𝒆 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆𝒔 ∶

𝑫𝒆𝒖𝒙 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆𝒔 𝒔𝒐𝒏𝒕 𝒅𝒊𝒕𝒆𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 ∶
∀(𝒙, 𝒚) ∈ ℝ𝟐 , 𝒇𝑿,𝒀 (𝒙, 𝒚) = 𝒇𝑿 (𝒙)𝒇𝒀 (𝒚)

𝓒– 𝟓 • 𝑰𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒄𝒆 𝒅’𝒖𝒏𝒆 𝒇𝒂𝒎𝒊𝒍𝒍𝒆 𝒇𝒊𝒏𝒊𝒆 𝒅𝒆 𝒗. 𝒂. ∶

𝑳𝒆𝒔 𝒎 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆𝒔 𝑿𝟏 , 𝑿𝟐 , … 𝑿𝒎 𝒔𝒐𝒏𝒕 𝒅𝒊𝒕𝒆𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 , 𝒔𝒊 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕𝒆𝒔
𝒑𝒂𝒓𝒕𝒊𝒆𝒔 𝑨𝟏 , 𝑨𝟐 , … 𝑨𝒎 𝒅𝒆 ℝ , 𝒍𝒆𝒔 é𝒗é𝒏𝒆𝒎𝒆𝒏𝒕𝒔{𝑿𝟏 ∈ 𝑨𝟏 }, {𝑿𝟐 ∈ 𝑨𝟐 }, … {𝑿𝒎 ∈ 𝑨𝒎 } 𝒔𝒐𝒏𝒕
𝒎

𝒎𝒖𝒕𝒖𝒆𝒍𝒍𝒆𝒎𝒆𝒏𝒕𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒔 ∶ 𝒇𝑿𝟏 ,𝑿𝟐 ,…𝑿𝒎 (𝒙𝟏𝒊 , 𝒙𝟐𝒊 , … , 𝒙𝒎𝒊 ) = ∏ 𝒇𝑿𝒌 (𝒙𝒌𝒊 )
𝒌=𝟏

𝓒– 𝟔 • 𝑰𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒄𝒆 𝒅’𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. ∶

303 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝑼𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 (𝑿𝒏 )𝒏∈ℕ 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒆𝒔𝒕 𝒅𝒊𝒕𝒆 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆 , 𝒔𝒊 𝒕𝒐𝒖𝒕𝒆
𝒔𝒐𝒖𝒔‑𝒔𝒖𝒊𝒕𝒆 𝒇𝒊𝒏𝒊𝒆 𝒆𝒔𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆 𝒂𝒖 𝒔𝒆𝒏𝒔 𝒅𝒆 𝓒– 𝟓

𝓒– 𝟕 • 𝑰𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒄𝒆 𝒅𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏𝒔 𝒅𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 ∶

𝑺𝒊 𝑿 𝒆𝒕 𝒀 𝒔𝒐𝒏𝒕 𝒅𝒆𝒖𝒙 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒆𝒕 𝒇 𝒆𝒕 𝒈 𝒅𝒆𝒖𝒙 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏𝒔 𝒕𝒆𝒍𝒍𝒆𝒔
𝒒𝒖𝒆 𝒍𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒇(𝑿)𝒆𝒕 𝒈(𝒀) 𝒔𝒐𝒏𝒕 𝒊𝒏𝒕é𝒈𝒓𝒂𝒃𝒍𝒆𝒔, 𝒂𝒍𝒐𝒓𝒔, 𝒐𝒏 𝒂 ∶
𝑬[𝒇(𝑿)𝒈(𝒀 )] = 𝑬[𝒇(𝑿)]𝑬[𝒈(𝒀 )] 𝒆𝒕 𝒍𝒆𝒔 𝒗. 𝒂. 𝒇(𝑿) 𝒆𝒕 𝒈(𝒀) 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔
𝓒– 𝟖 • 𝑺𝒐𝒎𝒎𝒆 𝒅𝒆 𝒅𝒆𝒖𝒙 𝒗. 𝒂. ∶ 𝑺𝒊 𝒍𝒆 𝒄𝒐𝒖𝒑𝒍𝒆 (𝑿, 𝒀) 𝒂𝒅𝒎𝒆𝒕 𝒑𝒐𝒖𝒓 𝒅𝒆𝒏𝒔𝒊𝒕é
𝒄𝒐𝒏𝒋𝒐𝒊𝒏𝒕𝒆 𝒇𝑿,𝒀 , 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒁 = 𝑿 + 𝒀 𝒂𝒖𝒓𝒂 é𝒈𝒂𝒍𝒆𝒎𝒆𝒏𝒕 𝒖𝒏𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒈 ∶

𝒈(𝒛) = ∫ 𝒇𝑿,𝒀 (𝒙, 𝒛 − 𝒙)𝒅𝒙 = ∫ 𝒇𝑿,𝒀 (𝒛 − 𝒚, 𝒚)𝒅𝒚

ℝ ℝ

 𝑺𝒊, 𝒆𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓, 𝑿 𝒆𝒕 𝒀 𝒔𝒐𝒏𝒕 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 , 𝒂𝒍𝒐𝒓𝒔 𝒍𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒏𝒐𝒕é𝒆 𝒈 𝒅𝒆

𝒁 = 𝑿 + 𝒀 𝒆𝒔𝒕 𝒍𝒆 𝒑𝒓𝒐𝒅𝒖𝒊𝒕 𝒄𝒐𝒏𝒗𝒐𝒍𝒖𝒕𝒊𝒐𝒏 𝒅𝒆 𝒇𝑿 𝒆𝒕 𝒇𝒀 ∶

𝒈(𝒛) = ∫ 𝒇𝑿 (𝒙)𝒇𝒀 (𝒛 − 𝒙)𝒅𝒙 = ∫ 𝒇𝑿 (𝒛 − 𝒚)𝒇𝒀 (𝒚)𝒅𝒚

ℝ ℝ

𝓒– 𝟗 • 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅′ 𝒖𝒏 𝒄𝒐𝒖𝒑𝒍𝒆 𝒅𝒆 𝒗. 𝒂. 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆𝒔 ∶

𝒙 𝒚
𝑭𝑿,𝒀 (𝒙, 𝒚) = 𝑷(𝑿 ≤ 𝒙, 𝒀 ≤ 𝒚) = ∫ ∫ 𝒇𝑿,𝒀 (𝒖, 𝒗)𝒅𝒖𝒅𝒗
−∞ −∞

𝝏𝟐 𝑭𝑿,𝒀 (𝒙, 𝒚)
 = 𝒇𝑿,𝒀 (𝒙, 𝒚)  ∀(𝒙, 𝒚) , 𝟎 ≤ 𝑭𝑿,𝒀 (𝒙, 𝒚) ≤ 𝟏
𝝏𝒚𝝏𝒙
 𝑭𝑿,𝒀 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒄𝒓𝒐𝒊𝒔𝒔𝒂𝒏𝒕𝒆
 𝐥𝐢𝐦 𝑭𝑿,𝒀 (𝒙, 𝒚) = 𝐥𝐢𝐦 𝑭𝑿,𝒀 (𝒙, 𝒚) = 𝟎 𝒐𝒖 𝒆𝒏𝒄𝒐𝒓𝒆 𝑭𝑿,𝒀 (−∞, 𝒚) = 𝑭𝑿,𝒀 (𝒙, −∞) = 𝟎
𝒙→−∞ 𝒚→−∞

 𝒍𝒊𝒎 𝒍𝒊𝒎 𝑭𝑿,𝒀 (𝒙, 𝒚) = 𝟏 𝒐𝒖 𝒆𝒏𝒄𝒐𝒓𝒆 𝑭𝑿,𝒀 (+∞, +∞) = 𝟏

𝒙→+∞ 𝒚→+∞

𝒃 𝒅
 𝑷(𝒂 < 𝑿 ≤ 𝒃, 𝒄 < 𝒀 ≤ 𝒅) = 𝑭𝑿,𝒀 (𝒃, 𝒅) + 𝑭𝑿,𝒀 (𝒂, 𝒄) − 𝑭𝑿,𝒀 (𝒂, 𝒅) − 𝑭𝑿,𝒀 (𝒃, 𝒄) = ∫ ∫ 𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚
𝒂 𝒄

 𝑭𝑿 = 𝑭𝒀 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 𝒍𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝑿 𝒆𝒕 𝒀 𝒐𝒏𝒕 𝒎ê𝒎𝒆 𝒍𝒐𝒊

𝓒– 𝟏𝟎 • 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏𝒔 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏𝒔 𝒎𝒂𝒓𝒈𝒊𝒏𝒂𝒍𝒆𝒔 ∶
𝒙
 𝑭𝑿 (𝒙) = 𝒍𝒊𝒎 𝑭𝑿,𝒀 (𝒙, 𝒚) = 𝑭𝑿,𝒀 (𝒙, +∞) = 𝑷𝑿 (𝑿 ≤ 𝒙) = ∫ 𝒇𝑿 (𝒖)𝒅𝒖
𝒚→+∞ −∞
𝒙
 𝑭𝒀 (𝒚) = 𝒍𝒊𝒎 𝑭𝑿,𝒀 (𝒙, 𝒚) = 𝑭𝑿,𝒀 (+∞, 𝒚) = 𝑷𝑿 (𝒀 ≤ 𝒚) = ∫ 𝒇𝒀 (𝒗)𝒅𝒗
𝒙→+∞ −∞

 𝑺𝒊 𝑿 𝒆𝒕 𝒀 𝒔𝒐𝒏𝒕 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 , 𝒂𝒍𝒐𝒓𝒔 𝑭𝑿,𝒀 (𝒙, 𝒚) = 𝑭𝑿 (𝒙)𝑭𝒀 (𝒚)

304 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝓒– 𝟏𝟏 • 𝑪𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏𝒏𝒆𝒎𝒆𝒏𝒕 𝒐𝒖 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏𝒔 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏𝒏𝒆𝒍𝒍𝒆𝒔 ∶
𝒇𝑿,𝒀 (𝒙, 𝒚) 𝒇𝑿,𝒀 (𝒙, 𝒚)
 𝒇(𝒀|𝑿 = 𝒙) (𝒚) =  𝒇(𝑿|𝒀 = 𝒚) (𝒙) =
𝒇𝑿 (𝒙) 𝒇𝒀 (𝒚)
+∞ +∞
 ∫ 𝒇(𝒀|𝑿 = 𝒙) (𝒚)𝒅𝒚 = 𝟏  ∫ 𝒇(𝑿|𝒀 = 𝒚) (𝒙)𝒅𝒙 = 𝟏
−∞ −∞

 𝑺𝒊 𝒍𝒂 𝒓𝒆𝒑𝒓é𝒔𝒆𝒏𝒕𝒂𝒕𝒊𝒐𝒏 𝒈𝒓𝒂𝒑𝒉𝒊𝒒𝒖𝒆 𝒅𝒆 𝑿(𝝎) × 𝒀(𝝎) 𝒏′ 𝒆𝒔𝒕 𝒑𝒂𝒔 𝒖𝒏 𝒓𝒆𝒄𝒕𝒂𝒏𝒈𝒍𝒆 𝒂𝒍𝒐𝒓𝒔 𝑿 𝒆𝒕 𝒀

𝒏𝒆 𝒔𝒐𝒏𝒕 𝒑𝒂𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔

𝓒– 𝟏𝟐 • 𝑳’𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 𝒅’𝒖𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝝋 𝒅𝒆 (𝑿, 𝒀) ∶

 𝑳′ 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒅𝒖 𝒄𝒐𝒖𝒑𝒍𝒆 (𝑿, 𝒀) 𝒆𝒔𝒕 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒔𝒊 𝑬(𝑿) 𝒆𝒕 𝑬(𝒀) 𝒆𝒙𝒊𝒔𝒕𝒆𝒏𝒕.
𝑶𝒏 𝒂 𝒂𝒍𝒐𝒓𝒔 ∶ 𝑬(𝑿, 𝒀) = (𝑬(𝑿), 𝑬(𝒀))

 𝑺𝒊 (𝑿, 𝒀) 𝒆𝒔𝒕 𝒖𝒏 𝒄𝒐𝒖𝒑𝒍𝒆 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆, 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝝋 ∶ ℝ𝟐 ⟶ ℝ

+∞ +∞
𝒕𝒆𝒍𝒍𝒆 𝒒𝒖𝒆 ∬ |𝝋(𝒙, 𝒚)|𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 , 𝒂𝒍𝒐𝒓𝒔 ∶ 𝑬[𝝋(𝑿, 𝒀)] = ∫ ∫ 𝝋(𝒙, 𝒚)𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚
ℝ𝟐 −∞ −∞

𝑪𝒆 𝒒𝒖𝒊 𝒏𝒐𝒖𝒔 𝒑𝒆𝒓𝒎𝒆𝒕 𝒅′ é𝒄𝒓𝒊𝒓𝒆 ∶

+∞ +∞ +∞ +∞
𝓪 • 𝑬(𝑿) = ∫ ∫ 𝒙𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚 𝓫 • 𝑬(𝒀) = ∫ ∫ 𝒚𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚
−∞ −∞ −∞ −∞
+∞ +∞
𝓬 • 𝑬(𝑿𝒀) = ∫ ∫ 𝒙𝒚𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚 𝓭 • 𝑬(𝑨𝑿𝑩 + 𝑪) = 𝑨𝑬(𝑿)𝑩 + 𝑪
−∞ −∞

𝓒– 𝟏𝟑 • 𝑬𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏𝒏𝒆𝒍𝒍𝒆 ∶
𝑳′ 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒓𝒈𝒊𝒏𝒂𝒍𝒆 𝒆𝒔𝒕 𝒍′𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒅𝒆𝒔 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆𝒔 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏𝒏𝒆𝒍𝒍𝒆𝒔

𝓪 • 𝑳′ 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏𝒏𝒆𝒍𝒍𝒆 𝑬(𝒀|𝑿) 𝒅𝒆 𝒀 𝒔𝒂𝒄𝒉𝒂𝒏𝒕 𝑿 𝒐ù 𝑿 𝒆𝒕 𝒀 𝒔𝒐𝒏𝒕 𝒅𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔

+∞
𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 ∶ 𝑬(𝒀|𝑿 = 𝒙) = ∫ 𝒚𝒇(𝒀|𝑿 = 𝒙) (𝒚)𝒅𝒚
−∞
+∞
⇒ 𝑬(𝒀) = 𝑬(𝑬(𝒀|𝑿 = 𝒙)) = ∫ 𝑬(𝒀|𝑿 = 𝒙)𝒇𝑿 (𝒙)𝒅𝒙
−∞

𝓫 • 𝑳′ 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏𝒏𝒆𝒍𝒍𝒆 𝑬(𝑿|𝒀) 𝒅𝒆 𝑿 𝒔𝒂𝒄𝒉𝒂𝒏𝒕 𝒀 𝒐ù 𝑿 𝒆𝒕 𝒀 𝒔𝒐𝒏𝒕 𝒅𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔

+∞
𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆𝒔 𝑬(𝑿|𝒀 = 𝒚) = ∫ 𝒙𝒇(𝑿|𝒀 = 𝒚) (𝒙)𝒅𝒙
−∞
+∞
⇒ 𝑬(𝑿) = 𝑬(𝑬(𝑿|𝒀 = 𝒚)) = ∫ 𝑬(𝑿|𝒀 = 𝒚)𝒇𝒀 (𝒚)𝒅𝒚
−∞

𝓬 • 𝑷𝒐𝒖𝒓 𝒕𝒐𝒖𝒕 𝒂 ∈ ℝ , 𝒃 ∈ ℝ , 𝑬(𝒂𝒀𝟏 + 𝒃𝒀𝟐 |𝑿) = 𝒂𝑬(𝒀𝟏 |𝑿) + 𝒃𝑬(𝒀𝟐 |𝑿)

𝓭 • 𝑺𝒊 𝑿 𝒆𝒕 𝒀 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 , 𝒂𝒍𝒐𝒓𝒔 𝑬(𝒀|𝑿) = 𝑬(𝒀)
𝓮 • 𝑺𝒊 𝑿 𝒆𝒕 𝒀 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 , 𝒂𝒍𝒐𝒓𝒔 𝑬(𝑿|𝒀) = 𝑬(𝑿)

305 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝓯 • (𝑰𝒏é𝒈𝒂𝒍𝒊𝒕é 𝒅𝒆 𝑪𝒂𝒖𝒄𝒉𝒚 𝑺𝒄𝒉𝒘𝒂𝒓𝒛) : |𝑬(𝑿𝒀)| ≤ 𝑬(|𝑿𝒀|) ≤ √𝑬(𝑿𝟐 )𝑬(𝒀𝟐 )

𝓒– 𝟏𝟒 • 𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆‑𝑪𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶
𝑺𝒐𝒊𝒆𝒏𝒕 𝑿 𝒆𝒕 𝒀 𝒅𝒆𝒖𝒙 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒔𝒖𝒓 (𝜴, 𝓕, 𝓟). 𝑶𝒏 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒅𝒆 𝑿 𝒆𝒕 𝒀 ,
𝒆𝒕 𝒍′ 𝒐𝒏 𝒏𝒐𝒕𝒆 𝑪𝒐𝒗(𝑿, 𝒀) , 𝒍𝒆 𝒓é𝒆𝒍 ∶
+∞ +∞
𝑪𝒐𝒗(𝑿, 𝒀) = 𝑬[(𝑿 − 𝑬(𝑿))(𝒀 − 𝑬(𝒀))] = ∫ ∫ (𝒙 − 𝑬(𝑿))(𝒚 − 𝑬(𝒀))𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚
−∞ −∞

 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 ∶ 𝑺𝒐𝒊𝒆𝒏𝒕 𝑿 , 𝒀 𝒆𝒕 𝒁 𝒅𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒔𝒖𝒓 (𝜴, 𝓕, 𝓟) 𝒆𝒕 𝒂, 𝒃, 𝒄, 𝒅 ∈ ℝ

𝓪 • 𝑪𝒐𝒗(𝑿, 𝑿) = 𝑽𝒂𝒓(𝑿) ≥ 𝟎 𝓫 • 𝑪𝒐𝒗(𝑿, 𝒀) = 𝑪𝒐𝒗(𝒀, 𝑿)
𝓬 • 𝑪𝒐𝒗(𝒂𝑿, 𝒀) = 𝑪𝒐𝒗(𝑿, 𝒂𝒀) = 𝒂𝑪𝒐𝒗(𝑿, 𝒀) 𝓭 • 𝑪𝒐𝒗(𝒂𝑿 + 𝒃𝒀, 𝒁) = 𝒂𝑪𝒐𝒗(𝑿, 𝒁) + 𝒃𝑪𝒐𝒗(𝑿, 𝒁)
𝓮 • 𝑪𝒐𝒗(𝑿, 𝒂𝒀 + 𝒃𝒁) = 𝒂𝑪𝒐𝒗(𝑿, 𝒀) + 𝒃𝑪𝒐𝒗(𝑿, 𝒁) 𝓯 • 𝑪𝒐𝒗(𝒂𝑿 + 𝒃, 𝒄𝒀 + 𝒅) = 𝒂𝒄𝑪𝒐𝒗(𝑿, 𝒀)
𝓰 • 𝑪𝒐𝒗(𝑿, 𝒃) = 𝑪𝒐𝒗(𝒂, 𝒀) = 𝟎 𝓱 • 𝑪𝒐𝒗(𝑿, 𝒀) = 𝑬(𝑿𝒀) − 𝑬(𝑿)𝑬(𝒀)
𝒏 𝒎 𝒏 𝒎

𝓲 • 𝑪𝒐𝒗 (∑ 𝑿𝒊 , ∑ 𝒀𝒋 ) = ∑ ∑ 𝑪𝒐𝒗(𝑿𝒊 , 𝒀𝒋 )
𝒊=𝟏 𝒋=𝟏 𝒊=𝟏 𝒋=𝟏

+∞ +∞
𝟐 𝟐
𝓳 • 𝑽𝒂𝒓 (𝑿) = 𝑬 [(𝑿 − 𝑬(𝑿)) ] = ∫ ∫ (𝒙 − 𝑬(𝑿)) 𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚
−∞ −∞
+∞ +∞
𝟐 𝟐
𝓴 • 𝑽𝒂𝒓 (𝒀) = 𝑬 [(𝒀 − 𝑬(𝒀)) ] = ∫ ∫ (𝒚 − 𝑬(𝒀)) 𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚
−∞ −∞

𝓵 • 𝑽(𝒂𝑿 + 𝒃𝒀) = 𝒂𝟐 𝑽(𝑿) + 𝒃𝟐 𝑽(𝒀) + 𝟐𝒂𝒃𝑪𝒐𝒗(𝑿, 𝒀)

𝟏
𝓶 • 𝑪𝒐𝒗(𝑿, 𝒀) = [𝑽𝒂𝒓(𝑿 + 𝒀) − 𝑽𝒂𝒓(𝑿) − 𝑽𝒂𝒓(𝒀)]
𝟐
𝓷 • 𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆𝒔 𝑪𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏𝒏𝒆𝒍𝒍𝒆𝒔 :
𝑳𝒂 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒎𝒂𝒓𝒈𝒊𝒏𝒂𝒍𝒆 𝒆𝒔𝒕 𝒍′𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒅𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆𝒔 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏𝒏𝒆𝒍𝒍𝒆𝒔 + 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒅𝒆𝒔 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆𝒔 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏𝒏𝒆𝒍𝒍𝒆𝒔

 𝑽(𝑿) = 𝑬 (𝑽(𝑿 = 𝒙𝒊 |𝒀 = 𝒚𝒋 )) + 𝑽 (𝑬(𝑿 = 𝒙𝒊 |𝒀 = 𝒚𝒋 ))

 𝑽(𝒀) = 𝑬 (𝑽(𝒀 = 𝒚𝒋 |𝑿 = 𝒙𝒊 )) + 𝑽 (𝑬(𝒀 = 𝒚𝒋 |𝑿 = 𝒙𝒊 ))

𝓸 • (𝑰𝒏é𝒈𝒂𝒍𝒊𝒕é 𝒅𝒆 𝑪𝒂𝒖𝒄𝒉𝒚 𝑺𝒄𝒉𝒘𝒂𝒓𝒛) : |𝑪𝒐𝒗(𝑿, 𝒀)| ≤ √𝑽𝒂𝒓(𝑿)𝑽𝒂𝒓(𝒀)

𝓹 • 𝑴𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆 𝑪𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑺𝒐𝒊𝒆𝒏𝒕 𝑿 𝒆𝒕 𝒀 𝒅𝒆𝒖𝒙 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒔𝒖𝒓 (𝛀, 𝓕, 𝓟)
𝑳𝒐𝒓𝒔𝒒𝒖𝒆 𝑽𝒂𝒓(𝑿), 𝑽𝒂𝒓(𝒀) 𝒆𝒕 𝑪𝒐𝒗(𝑿, 𝒀)𝒆𝒙𝒊𝒔𝒕𝒆𝒏𝒕 , 𝒍𝒂 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒅𝒖 𝒄𝒐𝒖𝒑𝒍𝒆 (𝑿, 𝒀)
𝑽𝒂𝒓(𝑿) 𝑪𝒐𝒗(𝑿, 𝒀)
𝒆𝒔𝒕 𝒍𝒂 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 ∶ 𝑪 = ( )
𝑪𝒐𝒗(𝑿, 𝒀) 𝑽𝒂𝒓(𝒀)
𝑼𝒏𝒆 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝑪 𝒆𝒔𝒕 𝒕𝒐𝒖𝒋𝒐𝒖𝒓𝒔 𝒖𝒏𝒆 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 𝒔𝒚𝒎é𝒕𝒓𝒊𝒒𝒖𝒆 𝒆𝒕 𝒑𝒐𝒔𝒊𝒕𝒊𝒗𝒆

306 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒏 𝒏

𝓺 • 𝑽𝒂𝒓 (∑ 𝑿𝒊 ) = ∑ 𝑽𝒂𝒓(𝑿𝒊 ) + 𝟐 ∑ 𝑪𝒐𝒗(𝑿𝒊 , 𝑿𝒋 )

𝒊=𝟏 𝒊=𝟏 𝟏≤𝒊<𝑗≤𝑛

 𝑪𝒂𝒔 𝒅𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 ∶

𝑺𝒐𝒊𝒆𝒏𝒕 𝑿 𝒆𝒕 𝒀 𝒅𝒆𝒖𝒙 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒔𝒖𝒓 (𝛀, 𝓕, 𝓟) 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 . 𝑶𝒏 𝒂 𝒂𝒍𝒐𝒓𝒔 ∶
⇒ 𝑬(𝑿𝒀) = 𝑬(𝑿)𝑬(𝒀)
𝑳𝒂 𝒓é𝒄𝒊𝒑𝒓𝒐𝒒𝒖𝒆 𝒅𝒆 𝒄𝒆 𝒓é𝒔𝒖𝒍𝒕𝒂𝒕 𝒆𝒔𝒕 𝒇𝒂𝒖𝒔𝒔𝒆

⇒ 𝑪𝒐𝒗(𝑿, 𝒀) = 𝟎
𝑳𝒂 𝒓é𝒄𝒊𝒑𝒓𝒐𝒒𝒖𝒆 𝒅𝒆 𝒄𝒆 𝒓é𝒔𝒖𝒍𝒕𝒂𝒕 𝒆𝒔𝒕 𝒇𝒂𝒖𝒔𝒔𝒆

⇒ 𝑽𝒂𝒓(𝑿 ± 𝒀) = 𝑽𝒂𝒓(𝑿) + 𝑽𝒂𝒓(𝒀)

𝑳𝒂 𝒓é𝒄𝒊𝒑𝒓𝒐𝒒𝒖𝒆 𝒅𝒆 𝒄𝒆 𝒓é𝒔𝒖𝒍𝒕𝒂𝒕 𝒆𝒔𝒕 𝒇𝒂𝒖𝒔𝒔𝒆

𝟐 𝟐
⇒ 𝑽𝒂𝒓(𝑿𝒀) = 𝑽𝒂𝒓(𝑿)𝑽𝒂𝒓(𝒀) + (𝑬(𝑿)) 𝑽𝒂𝒓(𝒀) + (𝑬(𝒀)) 𝑽𝒂𝒓(𝑿)
𝑳𝒂 𝒓é𝒄𝒊𝒑𝒓𝒐𝒒𝒖𝒆 𝒅𝒆 𝒄𝒆 𝒓é𝒔𝒖𝒍𝒕𝒂𝒕 𝒆𝒔𝒕 𝒇𝒂𝒖𝒔𝒔𝒆

𝑬𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 , 𝒔𝒊 𝑬(𝑿) = 𝑬(𝒀) = 𝟎 , 𝒂𝒍𝒐𝒓𝒔 𝑽𝒂𝒓(𝑿𝒀) = 𝑽𝒂𝒓(𝑿)𝑽𝒂𝒓(𝒀)

 𝑹𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒔: 𝑨𝒕𝒕𝒆𝒏𝒕𝒊𝒐𝒏, 𝒊𝒍 𝒏𝒆 𝒔’𝒂𝒈𝒊𝒕 𝒑𝒂𝒔 𝒅’𝒖𝒏𝒆 é𝒒𝒖𝒊𝒗𝒂𝒍𝒆𝒏𝒄𝒆: 𝒊𝒍 𝒆𝒙𝒊𝒔𝒕𝒆 𝒅𝒆𝒔 𝒄𝒐𝒖𝒑𝒍𝒆𝒔(𝑿, 𝒀)
𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒅𝒐𝒏𝒕 𝒍𝒂 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒆𝒔𝒕 𝒏𝒖𝒍𝒍𝒆 𝒎𝒂𝒊𝒔 𝒒𝒖𝒊 𝒏𝒆 𝒔𝒐𝒏𝒕 𝒑𝒂𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔.
 𝑹𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒔 ∶ 𝑺𝒐𝒊𝒕 (𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 ) 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 . 𝑶𝒏 𝒂 𝒂𝒍𝒐𝒓𝒔 ∶
𝒏 𝒏 𝒏 𝒏

𝑽𝒂𝒓 (∑ 𝑿𝒊 ) = ∑ 𝑽𝒂𝒓(𝑿𝒊 ) 𝒆𝒕 𝑬 (∏ 𝑿𝒊 ) = ∏ 𝑬(𝑿𝒊 )

𝒊=𝟏 𝒊=𝟏 𝒊=𝟏 𝒊=𝟏

𝓻 • 𝑽𝒂𝒓(𝑨𝑼) = 𝑨𝑽𝒂𝒓(𝑼)𝑨′ , 𝑼 é𝒕𝒂𝒏𝒕 𝒖𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆

𝓼 • 𝑪𝒐𝒗(𝑨𝑼, 𝑩𝑽) = 𝑨𝑪𝒐𝒗(𝑼, 𝑽)𝑩′ , 𝑼 𝒆𝒕 𝑽 𝒅𝒆𝒖𝒙 𝒗𝒆𝒄𝒕𝒆𝒖𝒓𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔
𝓽 • 𝑺𝒐𝒊𝒆𝒏𝒕 𝑨 𝒖𝒏𝒆 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 𝒔𝒚𝒎é𝒕𝒓𝒊𝒒𝒖𝒆 , 𝑪 = 𝑬(𝑼) 𝒆𝒕 𝚺 = 𝑽𝒂𝒓(𝑼) ∶ 𝑬(𝑼′ 𝑨𝑼) = 𝑻𝒓(𝑨𝚺) + 𝑪′ 𝑨𝑪

𝓒– 𝟏𝟓 • 𝑪𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅𝒆 𝒄𝒐𝒓𝒓é𝒍𝒂𝒕𝒊𝒐𝒏 𝒍𝒊𝒏é𝒂𝒊𝒓𝒆 ∶

𝑺𝒐𝒊𝒆𝒏𝒕 𝑿 𝒆𝒕 𝒀 𝒅𝒆𝒖𝒙 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒔𝒖𝒓 (𝛀, 𝓕, 𝓟) 𝒕𝒆𝒍𝒍𝒆𝒔 𝒒𝒖𝒆 𝝈(𝑿) ≠ 𝟎 𝒆𝒕 𝝈(𝒀) ≠ 𝟎 .
𝑳𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅𝒆 𝒄𝒐𝒓𝒓é𝒍𝒂𝒕𝒊𝒐𝒏 𝒍𝒊𝒏é𝒂𝒊𝒓𝒆 𝒆𝒏𝒕𝒓𝒆 𝑿 𝒆𝒕 𝒀 𝒆𝒔𝒕 𝒍𝒆 𝒏𝒐𝒎𝒃𝒓𝒆 𝒓é𝒆𝒍 𝝆(𝑿, 𝒀) ,
𝑪𝒐𝒗(𝑿, 𝒀) 𝑪𝒐𝒗(𝑿, 𝒀)
𝒅é𝒇𝒊𝒏𝒊 𝒑𝒂𝒓 ∶ 𝝆(𝑿, 𝒀) = =
√𝑽𝒂𝒓(𝑿)𝑽𝒂𝒓(𝒀) 𝝈(𝑿)𝝈(𝒀)

 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 ∶ 𝑺𝒐𝒊𝒆𝒏𝒕 𝑿 𝒆𝒕 𝒀 𝒅𝒆𝒖𝒙 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒔𝒖𝒓 (𝛀, 𝓕, 𝓟) 𝒕𝒆𝒍𝒍𝒆𝒔 𝒒𝒖𝒆

𝝈(𝑿) ≠ 𝟎 𝒆𝒕 𝝈(𝒀) ≠ 𝟎 .
𝓪 • −𝟏 ≤ 𝝆(𝑿, 𝒀) ≤ 𝟏 𝓫 • 𝝆(𝑿, 𝒀) = 𝟏 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 𝒀 = 𝒂 + 𝒃𝑿 , 𝒂 ∈ ℝ 𝒆𝒕 𝒃 > 𝟎
𝓬 • 𝝆(𝑿, 𝒀) = −𝟏 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 𝒀 = 𝒂 + 𝒃𝑿 , 𝒂 ∈ ℝ 𝒆𝒕 𝒃 < 𝟎
𝓭 • 𝑺𝒊 𝑿 𝒆𝒕 𝒀 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 , 𝒂𝒍𝒐𝒓𝒔 𝝆(𝑿, 𝒀) = 𝟎
𝓮 • 𝑺𝒊 𝝀 , 𝝁 ∈ ℝ∗+ , 𝒂𝒍𝒐𝒓𝒔 𝝆(𝝀𝑿, 𝝁𝒀) = 𝝆(𝑿, 𝒀)
 𝑹𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒔 ∶

307 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
 𝑼𝒏 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅𝒆 𝒄𝒐𝒓𝒓é𝒍𝒂𝒕𝒊𝒐𝒏 𝒍𝒊𝒏é𝒂𝒊𝒓𝒆 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒈𝒓𝒂𝒏𝒅𝒆𝒖𝒓 𝒔𝒂𝒏𝒔
𝒅𝒊𝒎𝒆𝒏𝒔𝒊𝒐𝒏 𝒆𝒕 𝒔𝒂𝒏𝒔 𝒖𝒏𝒊𝒕é (𝒅𝒊𝒕𝒆 𝒂𝒖𝒔𝒔𝒊 𝒈𝒓𝒂𝒏𝒅𝒆𝒖𝒓 𝒔𝒄𝒂𝒍𝒂𝒊𝒓𝒆). 𝑳𝒆 𝒅𝒆𝒓𝒏𝒊𝒆𝒓 𝒑𝒐𝒊𝒏𝒕 𝒔𝒊𝒈𝒏𝒊𝒇𝒊𝒆
𝒒𝒖’𝒊𝒍 𝒆𝒔𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕 𝒅𝒆𝒔 𝒖𝒏𝒊𝒕é𝒔 𝒄𝒉𝒐𝒊𝒔𝒊𝒆𝒔 𝒑𝒐𝒖𝒓 𝑿 𝒆𝒕 𝒀
 𝑺𝒊 𝝆(𝑿, 𝒀) > 𝟎 , 𝒐𝒏 𝒅𝒊𝒕 𝒒𝒖𝒆 𝑿 𝒆𝒕 𝒀 𝒔𝒐𝒏𝒕 𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔 𝒑𝒐𝒔𝒊𝒕𝒊𝒗𝒆𝒎𝒆𝒏𝒕
 𝑺𝒊 𝝆(𝑿, 𝒀) < 𝟎 , 𝒐𝒏 𝒅𝒊𝒕 𝒒𝒖𝒆 𝑿 𝒆𝒕 𝒀 𝒔𝒐𝒏𝒕 𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔 𝒏é𝒈𝒂𝒕𝒊𝒗𝒆𝒎𝒆𝒏𝒕
 𝑺𝒊 𝝆(𝑿, 𝒀) = 𝟎 , 𝒐𝒏 𝒅𝒊𝒕 𝒒𝒖𝒆 𝑿 𝒆𝒕 𝒀 𝒏𝒆 𝒔𝒐𝒏𝒕 𝒑𝒂𝒔 𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔
 𝑼𝒏𝒆 𝒄𝒐𝒓𝒓é𝒍𝒂𝒕𝒊𝒐𝒏 𝒑𝒐𝒔𝒊𝒕𝒊𝒗𝒆 𝒔𝒊𝒈𝒏𝒊𝒇𝒊𝒆 𝒒𝒖𝒆 𝒀 𝒂 «𝒕𝒆𝒏𝒅𝒂𝒏𝒄𝒆 à 𝒂𝒖𝒈𝒎𝒆𝒏𝒕𝒆𝒓» 𝒒𝒖𝒂𝒏𝒅 𝑿
𝒂𝒖𝒈𝒎𝒆𝒏𝒕𝒆 (𝒆𝒕, 𝒄𝒆 𝒒𝒖𝒊 𝒓𝒆𝒗𝒊𝒆𝒏𝒕 𝒂𝒖 𝒎ê𝒎𝒆, 𝒒𝒖𝒆 𝒀 𝒂 «𝒕𝒆𝒏𝒅𝒂𝒏𝒄𝒆 à 𝒂𝒖𝒈𝒎𝒆𝒏𝒕𝒆𝒓»𝒒𝒖𝒂𝒏𝒅 𝑿 𝒂𝒖𝒈𝒎𝒆𝒏𝒕𝒆)
 𝑼𝒏𝒆 𝒄𝒐𝒓𝒓é𝒍𝒂𝒕𝒊𝒐𝒏 𝒏é𝒈𝒂𝒕𝒊𝒗𝒆 𝒔𝒊𝒈𝒏𝒊𝒇𝒊𝒆 𝒒𝒖𝒆 𝒀 𝒂 𝒕𝒆𝒏𝒅𝒂𝒏𝒄𝒆 à 𝒅𝒊𝒎𝒊𝒏𝒖𝒆𝒓 𝒒𝒖𝒂𝒏𝒅 𝑿 𝒂𝒖𝒈𝒎𝒆𝒏𝒕𝒆,
𝒖𝒏𝒆 𝒂𝒃𝒔𝒆𝒏𝒄𝒆 𝒅𝒆 𝒄𝒐𝒓𝒓é𝒍𝒂𝒕𝒊𝒐𝒏 𝒒𝒖’𝒖𝒏𝒆 𝒂𝒖𝒈𝒎𝒆𝒏𝒕𝒂𝒕𝒊𝒐𝒏 𝒅𝒆 𝑿 𝒏’𝒂 𝒑𝒂𝒔𝒅’𝒊𝒏𝒇𝒍𝒖𝒆𝒏𝒄𝒆 𝒔𝒖𝒓 𝒍𝒂 «𝒗𝒂𝒍𝒆𝒖𝒓
𝒎𝒐𝒚𝒆𝒏𝒏𝒆» 𝒅𝒆 𝒀.
 𝑼𝒏 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅𝒆 𝒄𝒐𝒓𝒓é𝒍𝒂𝒕𝒊𝒐𝒏 𝒍𝒊𝒏é𝒂𝒊𝒓𝒆 «𝒑𝒓𝒐𝒄𝒉𝒆 𝒅𝒆 𝟏» 𝒆𝒏 𝒗𝒂𝒍𝒆𝒖𝒓 𝒂𝒃𝒔𝒐𝒍𝒖𝒆 𝒔𝒊𝒈𝒏𝒊𝒇𝒊𝒆 𝒒𝒖𝒆 𝒀
𝒑𝒆𝒖𝒕 ê𝒕𝒓𝒆 «𝒃𝒊𝒆𝒏 𝒂𝒑𝒑𝒓𝒐𝒄𝒉é𝒆» 𝒑𝒂𝒓 𝒖𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒂𝒇𝒇𝒊𝒏𝒆 𝒅𝒆 𝑿, 𝒄𝒓𝒐𝒊𝒔𝒔𝒂𝒏𝒕𝒆 𝒔𝒊 𝒍𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒆𝒔𝒕
𝒑𝒐𝒔𝒊𝒕𝒊𝒇, 𝒅é𝒄𝒓𝒐𝒊𝒔𝒔𝒂𝒏𝒕𝒆 𝒔𝒊𝒏𝒐𝒏. 𝑪’𝒆𝒔𝒕 𝒖𝒏𝒆 𝒒𝒖𝒆𝒔𝒕𝒊𝒐𝒏 𝒄𝒆𝒏𝒕𝒓𝒂𝒍𝒆𝒆𝒏 𝒔𝒕𝒂𝒕𝒊𝒔𝒕𝒊𝒒𝒖𝒆𝒔 (𝒎𝒐𝒊𝒏𝒔 𝒆𝒏
𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é𝒔)
 𝑫𝒆𝒖𝒙 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒔𝒐𝒏𝒕 𝒏𝒐𝒏 𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔 (𝒍𝒊𝒏é𝒂𝒊𝒓𝒆𝒎𝒆𝒏𝒕)𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 𝒍𝒆𝒖𝒓 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆
𝒆𝒔𝒕 𝒏𝒖𝒍𝒍𝒆. 𝑪𝒆𝒍𝒂 𝒏𝒆 𝒔𝒊𝒈𝒏𝒊𝒇𝒊𝒆 𝒑𝒂𝒔 𝒏é𝒄𝒆𝒔𝒔𝒂𝒊𝒓𝒆𝒎𝒆𝒏𝒕 𝒒𝒖’𝒆𝒍𝒍𝒆𝒔 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒔𝒂𝒖𝒇 𝒅𝒂𝒏𝒔 𝒍𝒆
𝒄𝒂𝒔 𝒕𝒓è𝒔 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 𝒅𝒆 𝒍𝒂 𝒑𝒓𝒐𝒑𝒓𝒊é𝒕é 𝒒𝒖𝒊 𝒔𝒖𝒊𝒕 ∶

 𝑺𝒐𝒊𝒕 (𝑿, 𝒀) 𝒖𝒏 𝒄𝒐𝒖𝒑𝒍𝒆 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 ∶ 𝑿 ↝ 𝓝(𝒎𝑿 , 𝝈𝟐𝑿 ) 𝒆𝒕 𝒀 ↝ 𝓝(𝒎𝒀 , 𝝈𝟐𝒀 ) . 𝑶𝒏 𝒂 ∶

𝑪𝒐𝒗(𝑿, 𝒀) = 𝟎 ⇔ 𝝆(𝑿, 𝒀) = 𝟎 ⇔ 𝑿 𝒆𝒕 𝒀 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔

𝓒– 𝟏𝟔 • 𝑭𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 ∶

𝑶𝒏 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒕𝒓𝒂𝒏𝒔𝒇𝒐𝒓𝒎é𝒆 𝒅𝒆 𝑳𝒂𝒑𝒍𝒂𝒄𝒆 (𝒐𝒖 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔)𝒅𝒖 𝒗𝒆𝒄𝒕𝒆𝒖𝒓
𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 (𝑿, 𝒀) (𝒔𝒊 𝒆𝒍𝒍𝒆 𝒆𝒙𝒊𝒔𝒕𝒆), 𝒍𝒂𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 ℝ𝟐 𝒅𝒂𝒏𝒔 ℝ 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒑𝒐𝒖𝒓 (𝒕𝟏 , 𝒕𝟐 ) 𝒑𝒂𝒓 ∶
+∞ +∞
(𝒕𝟏 𝑿+𝒕𝟐 𝒀)
𝑴(𝑿,𝒀) (𝒕𝟏 , 𝒕𝟐 ) = 𝑬(𝒆 )=∫ ∫ 𝒆(𝒕𝟏 𝒙+𝒕𝟐 𝒚) 𝒇𝑿,𝒀 (𝒙, 𝒚)𝒅𝒙𝒅𝒚
−∞ −∞

 𝑷𝒓𝒐𝒑𝒓𝒊é𝒕é𝒔 ∶
𝓪 • 𝑴(𝑿,𝒀) (𝟎, 𝟎) = 𝟏 𝓫 • 𝑴(𝑿,𝒀) (𝒕𝟏 , 𝒕𝟐 ) ≥ 𝟎 𝓬 • 𝑴(𝑿,𝒀) (𝒕𝟏 , 𝟎) = 𝑴𝑿 (𝒕𝟏 )

𝝏𝒓+𝒔 𝑴(𝑿,𝒀) (𝒕𝟏 , 𝒕𝟐 )

𝓭 • 𝑴(𝑿,𝒀) (𝟎, 𝒕𝟐 ) = 𝑴𝒀 (𝒕𝟐 ) 𝓮 • 𝑬(𝑿𝒓 . 𝒀𝒔 ) = [ ] |(𝟎, 𝟎)
𝝏𝒕𝒓𝟏 𝝏𝒕𝒔𝟐

𝝏𝟐 𝑴(𝑿,𝒀) (𝒕𝟏 , 𝒕𝟐 ) 𝝏𝑴(𝑿,𝒀) (𝒕𝟏 , 𝒕𝟐 )

𝓯 • 𝑬(𝑿𝒀) = [ ] |(𝟎, 𝟎) 𝓰 • 𝑬(𝑿) = [ ] |(𝟎, 𝟎)
𝝏𝒕𝟏 𝝏𝒕𝟐 𝝏𝒕𝟏

308 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝝏𝑴(𝑿,𝒀) (𝒕𝟏 , 𝒕𝟐 )
𝓱 • 𝑬(𝒀) = [ ] |(𝟎, 𝟎)
𝝏𝒕𝟐

𝓲 • 𝑿 𝒆𝒕 𝒀 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 , 𝒂𝒍𝒐𝒓𝒔 ∶ 𝑴(𝑿,𝒀) (𝒕𝟏 , 𝒕𝟐 ) = 𝑴𝑿 (𝒕𝟏 )𝑴𝒀 (𝒕𝟐 )

𝒏

𝓳 • 𝑺𝒊 (𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 ) 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 , 𝒂𝒍𝒐𝒓𝒔 ∶ 𝑴(𝑿𝟏,𝑿𝟐,…,𝑿𝒏 ) (𝒕𝟏, 𝒕𝟐 , … , 𝒕𝒏 ) = ∏ 𝑴𝑿𝒊 (𝒕𝒊 )

𝒊=𝟏
𝒏

𝓴 • 𝑺𝒊 (𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 ) 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒆𝒕 𝒂𝒊 ∈ ℝ, 𝒂𝒍𝒐𝒓𝒔 ∶ 𝑴∑𝒏𝒊=𝟏 𝒂𝒊𝑿𝒊 (𝒕) = ∏ 𝑴𝑿𝒊 (𝒂𝒊 𝒕)

𝒊=𝟏

𝓒– 𝟏𝟕 • 𝑴𝒂𝒙𝒊𝒎𝒖𝒎 𝒆𝒕 𝒎𝒊𝒏𝒊𝒎𝒖𝒎 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆𝒔 𝒆𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 ∶

𝒔𝒐𝒊𝒕 (𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 ) 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔. 𝑶𝒏 𝒑𝒐𝒔𝒆 𝑴𝒏 = 𝐦𝐚𝐱
∗
(𝑿𝒊 )
𝒊∈ℕ

𝒆𝒕 𝒎𝒏 = 𝐦𝐢𝐧∗ (𝑿𝒊 ) . 𝑨𝒍𝒐𝒓𝒔 ∶

𝒊∈ℕ
𝒏 𝒏

𝓪 • 𝑭𝑴𝒏 (𝒕) = ∏ 𝑭𝑿𝒊 (𝒕) ⇒ 𝒇𝑴𝒏 (𝒕) = 𝒅 [∏ 𝑭𝑿𝒊 (𝒕)]⁄𝒅𝒕

𝒊=𝟏 𝒊=𝟏
𝒏 𝒏

𝓫 • 𝑭𝒎𝒏 (𝒕) = 𝟏 − ∏ (𝟏 − 𝑭𝑿𝒊 (𝒕)) ⇒ 𝒇𝑴𝒏 (𝒕) = −𝒅 [∏ (𝟏 − 𝑭𝑿𝒊 (𝒕))]⁄𝒅𝒕

𝒊=𝟏 𝒊=𝟏

𝓓– 𝟏 • 𝑬𝒙𝒆𝒎𝒑𝒍𝒆 𝒇𝒐𝒏𝒅𝒂𝒎𝒆𝒏𝒕𝒂𝒍 ∶
𝑪𝒐𝒏𝒔𝒊𝒅é𝒓𝒐𝒏𝒔 𝒏 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝑿𝟏 , … , 𝑿𝒏 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒆𝒕 𝒅𝒆 𝒍𝒐𝒊𝒔 𝒓𝒆𝒔𝒑𝒆𝒄𝒕𝒊𝒗𝒆𝒎𝒆𝒏𝒕

𝓝(𝒎𝟏 , 𝝈𝟐𝟏 ), … , 𝓝(𝒎𝒏 , 𝝈𝟐𝒏 ). 𝒍𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒄𝒐𝒏𝒋𝒐𝒊𝒏𝒕𝒆 𝒅𝒖 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝑿 = (𝑿𝟏 ⋯ 𝑿𝒏 ) ′

𝟏 𝟏 𝟏
𝒇𝑿 (𝒙𝟏 ⋯ 𝒙𝒏 ) = 𝒏 × 𝐞𝐱𝐩 {− (𝒙 − 𝒎)′ 𝚺𝐗−𝟏 (𝒙 − 𝒎)} , 𝒂𝒗𝒆𝒄 𝒙 = (𝒙𝟏 ⋯ 𝒙𝒏 ) ′ ,
(√𝟐𝝅) √|𝚺𝐗 | 𝟐

𝝈𝟐𝟏 𝟎 ⋯ 𝟎
𝒎 = 𝑬(𝑿) = (𝑬(𝑿𝟏 ) ⋯ 𝑬(𝑿𝒏 ))′ = (𝒎𝟏 ⋯ 𝒎 𝒏 ) ′ , 𝚺𝐗 = ( 𝟎 ⋱ ⋱ ⋮
) 𝑳𝒂 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆
⋮ ⋱ ⋱ 𝟎
𝟎 ⋯ 𝟎 𝝈𝟐𝒏
𝟏⁄𝝈𝟐𝟏 𝟎 ⋯ 𝟎
𝟎 ⋱ ⋱ ⋮
𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆𝒔 𝒅𝒖 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝑿 𝒆𝒕 𝚺𝐗−𝟏 =( )
⋮ ⋱ ⋱ 𝟎
𝟎 ⋯ 𝟎 𝟏⁄𝝈𝟐𝒏
𝒏

𝑶𝒏 𝒗é𝒓𝒊𝒇𝒊𝒆 𝒒𝒖𝒆 ∶ (𝒙 − 𝒎)′ 𝚺𝐗−𝟏 (𝒙 − 𝒎) = ∑ (𝒙𝒊 − 𝒎𝒊 )𝟐 ⁄𝝈𝟐𝒊

𝒊=𝟏

309 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝓓– 𝟐 • 𝑫é𝒇𝒊𝒏𝒊𝒕𝒊𝒐𝒏 ∶
𝑼𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝑿 = (𝑿𝟏 ⋯ 𝑿𝒏 )′ 𝒅𝒆 ℝ𝒏 𝒆𝒔𝒕 𝒅𝒊𝒕 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 𝒔𝒊, 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕
𝒏

𝝀 = (𝝀𝟏 ⋯ )′ 𝒏 ′
𝝀𝒏 𝒅𝒆 ℝ 𝒍𝒂 𝒗. 𝒂. 𝝀 𝑿 = ∑ 𝝀𝒊 𝑿𝒊 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒅𝒆 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆.
𝒊=𝟏

𝑺𝒊 𝒔𝒐𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒅𝒆𝒔 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆𝒔 𝒆𝒔𝒕 𝒎 𝒆𝒕 𝒔𝒂 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆𝒔 𝒆𝒔𝒕 𝚺𝐗 , 𝒐𝒏 𝒏𝒐𝒕𝒆 ∶

𝑿 ↝ 𝓝(𝒎, 𝚺𝐗 )
𝓪 • 𝑹𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒔 ∶
 𝑼𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒅𝒐𝒏𝒕 𝒕𝒐𝒖𝒕𝒆𝒔 𝒍𝒆𝒔 𝒄𝒐𝒎𝒑𝒐𝒔𝒂𝒏𝒕𝒆𝒔 𝒔𝒐𝒏𝒕 𝒅𝒆 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆, 𝒏’𝒆𝒔𝒕𝒑𝒂𝒔 𝒏é𝒄𝒆𝒔𝒔𝒂𝒊𝒓𝒆𝒎𝒆𝒏𝒕
𝒖𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏.
 𝑻𝒐𝒖𝒕 𝒔𝒐𝒖𝒔 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒅’𝒖𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 𝒆𝒔𝒕 𝒆𝒏𝒄𝒐𝒓𝒆 𝒖𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏.
 𝑶𝒏 𝒑𝒂𝒓𝒍𝒆 𝒅𝒖 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 𝒔𝒕𝒂𝒏𝒅𝒂𝒓𝒅 𝒆𝒏 𝒅𝒊𝒎𝒆𝒏𝒔𝒊𝒐𝒏 𝒏 𝒔𝒊 𝑬(𝑿) = 𝟎𝒏×𝟏 𝒆𝒕 𝚺𝑿 = 𝑰𝒏
𝓫 • 𝑷𝒓𝒐𝒑𝒐𝒔𝒊𝒕𝒊𝒐𝒏 𝟏 ∶ 𝑺𝒊 𝑿 𝒆𝒔𝒕 𝒖𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 𝒅𝒆 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒅′ 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆
𝒎 = (𝒎𝟏 ⋯ 𝒎𝒏 )′ 𝒆𝒕 𝒅𝒆 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆𝒔 𝚺𝐗 , 𝒂𝒍𝒐𝒓𝒔, 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕 𝝀 𝒅𝒂𝒏𝒔 ℝ𝒏 , 𝒍𝒂 𝒗. 𝒂.

𝝀′ 𝑿 ↝ 𝓝(𝝀′ 𝒎, 𝝀′ 𝚺𝐗 𝝀)
𝓬 • 𝑷𝒓𝒐𝒑𝒐𝒔𝒊𝒕𝒊𝒐𝒏 𝟐 ∶
 𝑺𝒐𝒊𝒕 𝑿 𝒖𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 𝒅𝒆 ℝ𝒏 𝒅′ 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆𝒎 𝒆𝒕 𝒅𝒆 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆𝒔 𝚺𝐗 .
𝑳𝒐𝒓𝒔𝒒𝒖𝒆 𝚺𝐗 𝒆𝒔𝒕 𝒊𝒏𝒗𝒆𝒓𝒔𝒊𝒃𝒍𝒆, 𝒍𝒆 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝑿 𝒆𝒔𝒕 𝒅𝒊𝒕 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 𝒏𝒐𝒏
𝟏 𝟏 𝟏
𝒅é𝒈é𝒏é𝒓é 𝒆𝒕 𝒂𝒅𝒎𝒆𝒕 𝒑𝒐𝒖𝒓 𝒅𝒆𝒏𝒔𝒊𝒕é ∶ 𝒇𝑿 (𝒙𝟏 ⋯ 𝒙𝒏 ) = 𝒏 × 𝐞𝐱𝐩 {− (𝒙 − 𝒎)′ 𝚺−𝟏
𝐗 (𝒙 − 𝒎)}
(√𝟐𝝅) √|𝚺𝐗 | 𝟐

𝑼𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 𝒅𝒆 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆𝒔 𝚺𝐗 𝒕𝒆𝒍𝒍𝒆 𝒒𝒖𝒆| 𝚺𝐗 | = 𝟎

(𝒊. 𝒆. 𝚺𝐗 𝒏𝒐𝒏 𝒊𝒏𝒗𝒆𝒓𝒔𝒊𝒃𝒍𝒆) 𝒆𝒕 𝒅𝒊𝒕 𝒅é𝒈é𝒏é𝒓é 𝒆𝒕 𝒏′ 𝒂𝒅𝒎𝒆𝒕 𝒑𝒂𝒔 𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é.
′
 𝑷𝒐𝒖𝒓 𝒖𝒏 𝒄𝒐𝒖𝒑𝒍𝒆 (𝑿, 𝒀) 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏, 𝒐𝒏 𝒂 ∶ 𝒎 = 𝑬(𝑿) = (𝑬(𝑿), 𝑬(𝒀)) = (𝒎𝑿 , 𝒎𝒀 )′

𝛔𝟐𝐗 𝐂𝐨𝐯(𝑿, 𝒀)
𝒆𝒕 𝚺𝐗 = ( ) 𝒂𝒗𝒆𝒄 |𝚺𝐗 | = 𝛔𝟐𝐗 𝛔𝟐𝐘 − 𝐂𝐨𝐯 𝟐 (𝑿, 𝒀) = 𝛔𝟐𝐗 𝛔𝟐𝐘 (𝟏 − 𝝆𝟐 (𝑿, 𝒀))
𝐂𝐨𝐯(𝑿, 𝒀) 𝛔𝟐𝐘

𝑰𝒍 𝒆𝒏 𝒓é𝒔𝒖𝒍𝒕𝒆 ∶ ① (𝝆(𝑿, 𝒀) = ±𝟏 ⇔ 𝒀 = 𝒂 + 𝒃𝑿) ⇒ (𝑿, 𝒀) 𝒅é𝒈é𝒏é𝒓é 𝒆𝒕 𝒏′ 𝒂𝒅𝒎𝒆𝒕 𝒑𝒂𝒔 𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é

② 𝝆(𝑿, 𝒀) ≠ ±𝟏 ⇒ (𝑿, 𝒀) 𝒏𝒐𝒏 𝒅é𝒈é𝒏é𝒓é 𝒆𝒕 𝒂𝒅𝒎𝒆𝒕 𝒑𝒂𝒔 𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒇(𝑿,𝒀) , 𝒂𝒗𝒆𝒄 ,

𝟏 𝟏 (𝒙 − 𝒎𝑿 )𝟐 (𝒚 − 𝒎𝒀 )𝟐 𝟐𝝆(𝒙 − 𝒎𝑿 )(𝒚 − 𝒎𝒀 )
𝒇(𝑿,𝒀) (𝒙, 𝒚) = 𝐞𝐱𝐩 {− [ + − ]}
𝟐(𝟏 − 𝝆𝟐 ) 𝛔𝟐𝐗 𝛔𝟐𝐘 𝛔𝑿 𝛔𝒀
𝟐𝝅√𝛔𝟐𝐗 𝛔𝟐𝐘 (𝟏 − 𝝆𝟐 )

𝓭 • 𝑻𝒓𝒂𝒏𝒔𝒇𝒐𝒓𝒎𝒂𝒕𝒊𝒐𝒏 𝒍𝒊𝒏é𝒂𝒊𝒓𝒆 𝒅’𝒖𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 ∶ 𝑳𝒂 𝒕𝒓𝒂𝒏𝒔𝒇𝒐𝒓𝒎é𝒆 𝒅′ 𝒖𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓

𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 𝒅𝒆 ℝ𝒏 𝒑𝒂𝒕 𝒖𝒏𝒆 𝒂𝒑𝒑𝒍𝒊𝒄𝒂𝒕𝒊𝒐𝒏 𝒍𝒊𝒏é𝒂𝒊𝒓𝒆 𝒅𝒆 ℝ𝒏 𝒗𝒆𝒓𝒔 ℝ𝒑 𝒆𝒔𝒕 𝒆𝒏𝒄𝒐𝒓𝒆 𝒖𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏

310 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝑿 ↝ 𝓝(𝒎, 𝚺𝐗 )
{ ⇒ 𝑨𝑿 ↝ 𝓝(𝑨𝒎, 𝑨𝚺𝐗 𝑨′ )
𝑨 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆 𝒕𝒂𝒊𝒍𝒍𝒆 (𝒑 × 𝒏)
𝓮 • 𝑽𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 𝒔𝒕𝒂𝒏𝒅𝒂𝒓𝒅 ∶ 𝑺𝒐𝒊𝒕 𝑿 ↝ 𝓝(𝒎, 𝚺𝐗 ) , 𝒖𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏
𝟏
−
𝒏𝒐𝒏 𝒅é𝒈é𝒏é𝒓é , 𝒂𝒍𝒐𝒓𝒔 𝚺𝐗 𝟐 (𝑿 − 𝒎) ↝ 𝓝(𝟎𝒏×𝟏 , 𝑰𝒏 ) ∶ 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 𝒔𝒕𝒂𝒏𝒅𝒂𝒓𝒅

𝓓– 𝟑 • 𝑰𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒄𝒆 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏𝒏𝒆𝒔 ∶

𝓪 • 𝑷𝒓𝒐𝒑𝒐𝒔𝒊𝒕𝒊𝒐𝒏 𝟏 ∶ 𝑺𝒐𝒊𝒕 𝑿 𝒖𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 𝒅𝒆 ℝ𝒏 . 𝑷𝒐𝒖𝒓 𝒒𝒖𝒆 𝒔𝒆𝒔 𝒄𝒐𝒎𝒑𝒐𝒔𝒂𝒏𝒕𝒆𝒔
𝑿𝟏 , … , 𝑿𝒏 𝒔𝒐𝒊𝒆𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔, 𝒊𝒍 𝒇𝒂𝒖𝒕 𝒆𝒕 𝒊𝒍 𝒔𝒖𝒇𝒇𝒊𝒕 𝒒𝒖𝒆 𝒍𝒂 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆𝒔 𝒔𝒐𝒊𝒕
𝒅𝒊𝒂𝒈𝒐𝒏𝒂𝒍𝒆
𝓫 • 𝑪𝒐𝒓𝒐𝒍𝒍𝒂𝒊𝒓𝒆 ∶ 𝑺𝒊 𝒍𝒆 𝒄𝒐𝒖𝒑𝒍𝒆 (𝑿, 𝒀) 𝒆𝒔𝒕 𝒖𝒏 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏, 𝒐𝒏 𝒂 ∶
𝑿 𝒆𝒕 𝒀 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 ⇔ 𝑪𝒐𝒗(𝑿, 𝒀) = 𝟎
𝓬 • 𝑷𝒓𝒐𝒑𝒐𝒔𝒊𝒕𝒊𝒐𝒏 𝟐 ∶ 𝑳𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅′ 𝒖𝒏𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏 𝒔𝒕𝒂𝒏𝒅𝒂𝒓𝒅 𝒆𝒏 𝒅𝒊𝒎𝒆𝒏𝒔𝒊𝒐𝒏 𝒏 𝒆𝒔𝒕 ∶
𝒏
𝟏 𝟏 𝟐
𝒇𝑿 (𝒙𝟏 ⋯ 𝒙𝒏 ) = 𝒏 𝒆𝒙𝒑 {− ∑ 𝒙𝒊 }
(√𝟐𝝅) 𝟐
𝒊=𝟏

311 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014

Exercice 1 : (5 points :1+1+1+1+1)

ÉNONCÉ
𝑶𝒏 𝒄𝒐𝒏𝒔𝒊𝒅è𝒓𝒆 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝐗 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝒄𝒆𝒏𝒕𝒓é𝒆 𝒓é𝒅𝒖𝒊𝒕𝒆 𝐗 ↝ 𝓝(𝟎, 𝟏) .
𝑶𝒏 𝒑𝒐𝒔𝒆 𝒀 = |𝑿|
1) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒆 𝒅𝒐𝒎𝒂𝒊𝒏𝒆 𝒅𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 𝒑𝒐𝒔𝒔𝒊𝒃𝒍𝒆𝒔 𝒅𝒆 𝒀
2) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝑮(𝒚) 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒀 𝒆𝒏 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝑭 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆
𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝑿
3) 𝑬𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒍𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒅𝒆 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒀
4) 𝑷𝒓𝒐𝒖𝒗𝒆𝒓 𝒒𝒖𝒆 𝑬(𝒀𝟐 ) 𝒆𝒔𝒕 é𝒈𝒂𝒍𝒆 à 𝒍’𝒖𝒏𝒊𝒕é
5) 𝑷𝒓𝒐𝒖𝒗𝒆𝒓 𝒒𝒖𝒆 𝒍𝒂 𝒎é𝒅𝒊𝒂𝒏𝒆 𝒅𝒆 𝒀 𝒆𝒔𝒕 𝒍𝒆 𝒕𝒓𝒐𝒊𝒔𝒊è𝒎𝒆 𝒒𝒖𝒂𝒓𝒕𝒊𝒍𝒆 𝒅𝒆 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝑿

Corrigé
1) 𝑿 ↝ 𝓝(𝟎, 𝟏) ⟹ 𝛀𝑿 = ]−∞, +∞[

𝒀 = |𝑿| ⟹ 𝛀𝒀 = [𝟎, +∞[

2) 𝑮(𝒚) = 𝑷(𝒀 ≤ 𝒚) = 𝑷(|𝑿| ≤ 𝒚) = 𝑷(−𝒚 ≤ 𝑿 ≤ 𝒚) = 𝑭(𝒚) − 𝑭(−𝒚)

𝑭(𝒚) = 𝑷(𝑿 ≤ 𝒚) 𝒂𝒖𝒕𝒓𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒕 𝑭(𝒚) 𝒆𝒔𝒕 𝒍𝒂 𝒇. 𝒓 𝒅𝒆 𝒍𝒂 𝒗. 𝒂 𝑿

𝒐𝒓 𝑭(−𝒚) = 𝟏 − 𝑭(𝒚) , 𝒅𝒐𝒏𝒄 𝑮(𝒚) = 𝑭(𝒚) − (𝟏 − 𝑭(𝒚))

𝒅′ 𝒐ù 𝑮(𝒚) = 𝟐𝑭(𝒚) − 𝟏
𝒅𝑮(𝒚) 𝒅𝑭(𝒚)
3) 𝑺𝒐𝒊𝒕 𝒈(𝒚) 𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂 𝒀, 𝒐𝒓 = 𝒈(𝒚) 𝒆𝒕 = 𝒇(𝒚)
𝒅𝒚 𝒅𝒚

𝒅𝑮(𝒚) 𝒅[𝟐𝑭(𝒚) − 𝟏] 𝒅𝑭(𝒚)

𝑮(𝒚) = 𝟐𝑭(𝒚) − 𝟏 ⟹ ∀ 𝒚 ∈ [𝟎, +∞[ = ⟹ ∀ 𝒚 ∈ [𝟎, +∞[ 𝒈(𝒚) = 𝟐 ( )
𝒅𝒚 𝒅𝒚 𝒅𝒚

⟹ ∀ 𝒚 ∈ [𝟎, +∞[ 𝒈(𝒚) = 𝟐𝒇(𝒚)

𝟐 −𝒚𝟐
√ , ∀ 𝒚 ∈ [𝟎, +∞[
𝒅′ 𝒐ù 𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂 𝒀 ∶ 𝒈(𝒚) = { 𝝅 𝒆
𝟐

𝟎 , 𝒂𝒊𝒍𝒍𝒆𝒖𝒓𝒔
𝟐) 𝟐) 𝟐)
4) 𝑬(𝒀 = 𝑬(|𝑿| = 𝑬(𝑿

312 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝑽(𝑿) = 𝟏
𝟐
𝒐𝒓 {𝑽(𝑿) = 𝑬(𝑿𝟐 ) − (𝑬(𝑿)) ⟹ 𝑬(𝑿𝟐 ) = 𝟏 𝑫′ 𝒐ù 𝑬(𝒀𝟐 ) = 𝟏
𝑬(𝑿) = 𝟎
𝟏
5) 𝑺𝒐𝒊𝒕 𝑴𝒆𝒀 𝒍𝒂 𝒗𝒂𝒍𝒆𝒖𝒓 𝒎é𝒅𝒊𝒂𝒏𝒆 𝒅𝒆 𝒍𝒂 𝒗. 𝒂 𝒀, 𝒄𝒆 𝒒𝒖𝒊 𝒅𝒐𝒏𝒏𝒆 𝑮(𝑴𝒆𝒀 ) = 𝟐

𝟏
𝒐𝒓 𝑮(𝒚) = 𝟐𝑭(𝒚) − 𝟏 𝒅𝒐𝒏𝒄 𝒆𝒏 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒊𝒆𝒓 𝑮(𝑴𝒆𝒀 ) = 𝟐𝑭(𝑴𝒆𝒀 ) − 𝟏 ⟺ 𝟐𝑭(𝑴𝒆𝒀 ) − 𝟏 =
𝟐
𝟑 𝟑
⟺ 𝟐𝑭(𝑴𝒆𝒀 ) = ⟺ 𝑭(𝑴𝒆𝒀 ) =
𝟐 𝟒
𝟑
𝑵𝒐𝒕𝒐𝒏𝒔 𝑸𝟑 𝑿 𝒍𝒆 𝒕𝒓𝒐𝒊𝒔𝒊è𝒎𝒆 𝒒𝒖𝒂𝒓𝒕𝒊𝒍𝒆 𝒅𝒆 𝒍𝒂 𝒗. 𝒂 𝑿, 𝒄𝒆 𝒒𝒖𝒊 𝒅𝒐𝒏𝒏𝒆 𝑭(𝑸𝟑 𝑿 ) =
𝟒
𝑪𝒐𝒎𝒎𝒆 𝒕𝒐𝒖𝒕𝒆 𝒇. 𝒓 , 𝑭 𝒓é𝒂𝒍𝒊𝒔𝒆 𝒖𝒏𝒆 𝒃𝒊𝒋𝒆𝒄𝒕𝒊𝒐𝒏 𝒔𝒖𝒓 𝜴𝑿 𝒑𝒂𝒓 𝒍𝒖𝒔𝒖𝒊𝒕𝒆 ∶ 𝑭(𝑸𝟑 𝑿 ) = 𝑮(𝑴𝒆𝒀 )

𝒅′ 𝒐ù 𝑸𝟑 𝑿 = 𝑴𝒆𝒀

Exercice 2 : (8 points : 1+1,5+1+1,5+1,5+1,5)

ÉNONCÉ
𝑳𝒆 𝒓𝒆𝒏𝒅𝒆𝒎𝒆𝒏𝒕 𝑿 𝒅’𝒖𝒏𝒆 𝒂𝒄𝒕𝒊𝒐𝒏 à 𝒍𝒂 𝒃𝒐𝒖𝒓𝒔𝒆 𝒅𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 𝒔𝒖𝒊𝒕 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝒅’𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆
𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 𝒎 𝒆𝒕 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝝈𝟐 : 𝑿 ↝ 𝓝(𝒎, 𝝈𝟐 ). 𝑶𝒏 𝒏𝒐𝒕𝒆 𝑭(𝒙)𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏
𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝒄𝒆𝒏𝒕𝒓é𝒆‑𝒓é𝒅𝒖𝒊𝒕𝒆.
1) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒗𝒂𝒍𝒆𝒖𝒓 𝒏𝒖𝒎é𝒓𝒊𝒒𝒖𝒆 𝒅𝒆 𝒍𝒂 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒑𝒐𝒖𝒓 𝒒𝒖𝒆 𝑿 𝒔𝒐𝒊𝒕 𝒄𝒐𝒎𝒑𝒓𝒊𝒔𝒆 𝒆𝒏𝒕𝒓𝒆
𝒎 − 𝟐𝝈 𝒆𝒕 𝒎 + 𝟐𝝈 : (𝒎 − 𝟐𝝈 ≤ 𝑿 ≤ 𝒎 + 𝟐𝝈)
2) 𝑨𝒚𝒂𝒏𝒕 𝒏 𝒐𝒃𝒔𝒆𝒓𝒗𝒂𝒕𝒊𝒐𝒏𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒅𝒆 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝑿 𝒏𝒐𝒕é𝒆𝒔 𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 .
𝒏 𝒏
𝟏 𝟏 𝟐
𝑶𝒏 𝒑𝒐𝒔𝒆 : 𝑿 = ∑ 𝑿𝒊 𝒆𝒕 𝑺𝟐 = ∑(𝑿𝒊 − 𝑿)
𝒏 𝒏
𝒊=𝟏 𝒊=𝟏

a) 𝑸𝒖𝒆𝒍𝒍𝒆𝒔 𝒔𝒐𝒏𝒕 𝒍𝒆𝒔 𝒊𝒏𝒕𝒆𝒓𝒑𝒓é𝒕𝒂𝒕𝒊𝒐𝒏𝒔 é𝒄𝒐𝒏𝒐𝒎𝒊𝒒𝒖𝒆𝒔 𝒅𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝑿 𝒆𝒕 𝑺𝟐 ?

b) 𝑷𝒓𝒐𝒖𝒗𝒆𝒓 𝒒𝒖𝒆 𝑿 𝒆𝒔𝒕 𝒖𝒏 𝒆𝒔𝒕𝒊𝒎𝒂𝒕𝒆𝒖𝒓 𝒔𝒂𝒏𝒔 𝒃𝒊𝒂𝒊𝒔 𝒅𝒆 𝒎
c) 𝑫𝒂𝒏𝒔 𝒄𝒆𝒕𝒕𝒆 𝒒𝒖𝒆𝒔𝒕𝒊𝒐𝒏, 𝒐𝒏 𝒔𝒖𝒑𝒑𝒐𝒔𝒆 𝒒𝒖𝒆 𝒏 = 𝟐.
i. 𝑷𝒓𝒐𝒖𝒗𝒆𝒓 𝒒𝒖𝒆 𝑺𝟐 𝒔’é𝒄𝒓𝒊𝒕 𝒆𝒏 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 ( 𝑿𝟏 − 𝑿𝟐 )

313 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝟐
ii. 𝑽é𝒓𝒊𝒇𝒊𝒆𝒓 𝒒𝒖𝒆 𝑿 𝒆𝒕 𝑺 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔
iii. 𝑰𝒏𝒕𝒆𝒓𝒑𝒓é𝒕𝒆𝒓 𝒄𝒆 𝒅𝒆𝒓𝒏𝒊𝒆𝒓 𝒓é𝒔𝒖𝒍𝒕𝒂𝒕 𝒆𝒏 𝒕𝒆𝒓𝒎𝒆𝒔 𝒅𝒆 𝒓𝒆𝒍𝒂𝒕𝒊𝒐𝒏 𝒆𝒏𝒕𝒓𝒆 𝒍𝒂
𝒓𝒆𝒏𝒕𝒂𝒃𝒊𝒍𝒊𝒕é 𝒆𝒕 𝒍𝒆 𝒓𝒊𝒔𝒒𝒖𝒆.

Corrigé
𝑿−𝒎
1) 𝑷(𝒎 − 𝟐𝝈 ≤ 𝑿 ≤ 𝒎 + 𝟐𝝈) = 𝑷 (−𝟐 ≤ ≤ 𝟐)
𝝈

𝑿−𝒎
𝑨𝒊𝒏𝒔𝒊 𝑷(𝒎 − 𝟐𝝈 ≤ 𝑿 ≤ 𝒎 + 𝟐𝝈) = 𝑷(−𝟐 ≤ 𝒀 ≤ 𝟐)𝒐ù 𝒀 = ↝ 𝓝(𝟎, 𝟏), 𝒔𝒊 𝑿 ↝ 𝓝(𝒎, 𝝈𝟐 )
𝝈
𝑷(𝒎 − 𝟐𝝈 ≤ 𝑿 ≤ 𝒎 + 𝟐𝝈) = 𝐅(𝟐) − 𝐅(−𝟐) = 𝐅(𝟐) − (𝟏 − 𝐅(𝟐)) = 𝟐𝐅(𝟐) − 𝟏 = (𝟐 × 𝟎, 𝟗𝟕𝟕𝟐) − 𝟏

𝑷(𝒎 − 𝟐𝝈 ≤ 𝑿 ≤ 𝒎 + 𝟐𝝈) = 𝟗𝟓, 𝟒𝟒%

2)
𝟏
a) 𝑿 = 𝒏 ∑𝒏𝒊=𝟏 𝑿𝒊 ∶ 𝑳𝒆 𝒓𝒆𝒏𝒅𝒆𝒎𝒆𝒏𝒕 𝒎𝒐𝒚𝒆𝒏 𝒅𝒆𝒔 𝒂𝒄𝒕𝒊𝒐𝒏𝒔

𝒏
𝟐
𝟏 𝟐
𝑺 = ∑(𝑿𝒊 − 𝑿) ∶ 𝑼𝒏 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅𝒆 𝒓𝒊𝒔𝒒𝒖𝒆, 𝒎𝒆𝒔𝒖𝒓𝒂𝒏𝒕 𝒍𝒆𝒔 é𝒄𝒂𝒓𝒕𝒔 𝒄𝒐𝒏𝒔𝒕𝒂𝒕é𝒔 𝒆𝒏𝒕𝒓𝒆 𝒍𝒆𝒔
𝒏
𝒊=𝟏

𝒅𝒊𝒇𝒇é𝒓𝒆𝒏𝒕𝒔 𝒓𝒆𝒏𝒅𝒆𝒎𝒆𝒏𝒕𝒔 𝒆𝒕 𝒍𝒆 𝒓𝒆𝒏𝒅𝒆𝒎𝒆𝒏𝒕 𝒎𝒐𝒚𝒆𝒏.

𝟏 𝟏 𝟏 𝟏 𝟏
b) 𝑬(𝑿) = 𝑬 (𝒏 ∑𝒏𝒊=𝟏 𝑿𝒊 ) = 𝒏 𝑬(∑𝒏𝒊=𝟏 𝑿𝒊 ) = 𝒏 ∑𝒏𝒊=𝟏 𝑬(𝑿𝒊 ) = 𝒏 ∑𝒏𝒊=𝟏 𝒎 = 𝒏 (𝒏. 𝒎) = 𝒎

̂ = 𝑿 𝒆𝒔𝒕 𝒖𝒏 𝒆𝒔𝒕𝒊𝒎𝒂𝒕𝒆𝒖𝒓 𝒔𝒂𝒏𝒔 𝒃𝒊𝒂𝒊𝒔 𝒅𝒆 𝒎

𝒎
c)
𝟏 𝟐 𝟏 𝟐 𝟐 𝟏 𝑿𝟏 +𝑿𝟐
i. 𝑺𝟐 = 𝟐 ∑𝟐𝒊=𝟏(𝑿𝒊 − 𝑿) = 𝟐 [(𝑿𝟏 − 𝑿) + (𝑿𝟐 − 𝑿) ] . 𝑶𝒓 𝑿 = 𝟐 ∑𝟐𝒊=𝟏 𝑿𝒊 = 𝟐

𝟐 𝟐
𝟏 𝟐
𝑿𝟏 + 𝑿𝟐 𝑿𝟏 + 𝑿𝟐 𝟏 𝑿𝟏 − 𝑿𝟐 𝟐 𝑿𝟏 − 𝑿𝟐 𝟐
𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 𝑺 = [(𝑿𝟏 − ( )) + (𝑿𝟐 − ( )) ] = [( ) +( ) ]
𝟐 𝟐 𝟐 𝟐 𝟐 𝟐

𝟐
𝑿𝟏 − 𝑿𝟐 𝟐
𝑺 =( )
𝟐
ii.
𝟐
𝑿 ↝ 𝓝(𝒎, 𝝈𝟐 ) ; 𝒊 = 𝟏, 𝟐 𝑿𝟏 − 𝑿𝟐 𝝈
{ 𝒊 ⇒ (𝑿𝟏 − 𝑿𝟐 ) ↝ 𝓝(𝟎, 𝟐𝝈𝟐 ) ⇒ 𝑺 = ( ) ↝ 𝓝 (𝟎, )
𝑿𝟏 𝒆𝒕 𝑿𝟐 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝟐 𝟐

𝝈𝟐
𝒆𝒕 𝑿 ↝ 𝓝 (𝒎, ) 𝒅𝒐𝒏𝒄 𝑿 𝒆𝒕 𝑺 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒈𝒂𝒖𝒔𝒔𝒊𝒆𝒏𝒏𝒆𝒔
𝟐

𝑪𝒂𝒍𝒄𝒖𝒍𝒐𝒏𝒔 𝑪𝒐𝒗(𝑿, 𝑺) ∶
𝑿𝟏 + 𝑿𝟐 𝑿𝟏 − 𝑿𝟐 𝟏
𝑪𝒐𝒗(𝑿, 𝑺) = 𝑪𝒐𝒗 [( ),( )] = 𝑪𝒐𝒗[(𝑿𝟏 + 𝑿𝟐 ), (𝑿𝟏 − 𝑿𝟐 )]
𝟐 𝟐 𝟒

314 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
= 𝟏 𝟒 [𝑪𝒐𝒗[𝑿𝟏 , 𝟏 − 𝑿𝟐 + 𝑪𝒐𝒗[𝑿𝟐 , 𝟏 − 𝑿𝟐 )]]
⁄ (𝑿 )] (𝑿

= 𝟏⁄𝟒 [𝑪𝒐𝒗(𝑿𝟏 , 𝑿𝟏 ) + 𝑪𝒐𝒗(𝑿𝟏 , −𝑿𝟐 ) + 𝑪𝒐𝒗(𝑿𝟐 , 𝑿𝟏 ) + 𝑪𝒐𝒗(𝑿𝟐 , −𝑿𝟐 )]

= 𝟏⁄𝟒 [𝑽𝒂𝒓(𝑿𝟏 ) − 𝑪𝒐𝒗(𝑿𝟏 , 𝑿𝟐 ) + 𝑪𝒐𝒗(𝑿𝟏 , 𝑿𝟐 ) − 𝑪𝒐𝒗(𝑿𝟐 , 𝑿𝟐 )] = 𝟏⁄𝟒 [𝑽𝒂𝒓(𝑿𝟏 ) − 𝑽𝒂𝒓(𝑿𝟐 )]

𝑪𝒐𝒗(𝑿, 𝑺) = 𝟎 𝒆𝒕 𝑿 𝒆𝒕 𝑺 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒏𝒐𝒓𝒎𝒂𝒍𝒆𝒔 , 𝒅𝒐𝒏𝒄 𝑿 𝒆𝒕 𝑺 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔

𝑺𝒐𝒊𝒆𝒏𝒕 𝒇(𝑿 ) = 𝑿 𝒆𝒕 𝒈(𝑺) = 𝑺𝟐 , 𝑬[𝒇(𝑿 )] = 𝑬(𝑿 ) = 𝒎

𝟐
′
𝑿𝟏 − 𝑿𝟐 𝝈𝟐 𝑺 𝑺
𝑫 𝒂𝒖𝒕𝒓𝒆 𝒑𝒂𝒓𝒕 𝑺 = ( ) ↝ 𝓝 (𝟎, ) ⇒ ↝ 𝓝(𝟎, 𝟏) ⇒ ( ) ↝ 𝝌𝟐 (𝟏)
𝟐 𝟐 ⁄
𝝈 √𝟐 ⁄
𝝈 √𝟐
𝟐
𝑺 𝑺𝟐 𝝈𝟐
𝒂𝒊𝒏𝒔𝒊 𝑬 [( ) ] = 𝟏 ⇔ 𝑬 ( 𝟐 ) = 𝟏 ⇔ 𝑬(𝑺𝟐 ) =
𝝈⁄√𝟐 𝝈 ⁄𝟐 𝟐

𝝈𝟐
𝒆𝒕 𝑬[𝒈(𝑺)] = 𝑬(𝑺𝟐 ) = 𝒆𝒏 𝒆𝒇𝒇𝒆𝒕 (𝒍𝒆𝒔 𝒗. 𝒂. 𝒇(𝑿)𝒆𝒕 𝒈(𝒀) 𝒔𝒐𝒏𝒕 𝒊𝒏𝒕é𝒈𝒓𝒂𝒃𝒍𝒆𝒔)
𝟐

𝒅′ 𝒐ù 𝑿 𝒆𝒕 𝑺𝟐 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔

iii.

𝑺𝟐 𝒎𝒆𝒔𝒖𝒓𝒆 𝒍𝒆 𝒓𝒊𝒔𝒒𝒖𝒆 , 𝑿 é𝒕𝒂𝒏𝒕 𝒍𝒆 𝒓𝒆𝒏𝒅𝒆𝒎𝒆𝒏𝒕 𝒎𝒐𝒚𝒆𝒏 𝒅′ 𝒖𝒏 𝒑𝒐𝒓𝒕𝒆𝒇𝒆𝒖𝒊𝒍𝒍𝒆 é𝒒𝒖𝒊𝒑𝒐𝒏𝒅é𝒓é.

𝑿 𝒆𝒕 𝑺𝟐 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 , 𝒄𝒆𝒍𝒂 𝒊𝒎𝒑𝒍𝒊𝒒𝒖𝒆 𝒍𝒆 𝒓𝒆𝒏𝒅𝒆𝒎𝒆𝒏𝒕 𝒎𝒐𝒚𝒆𝒏 𝒅′ 𝒖𝒏 𝒑𝒐𝒓𝒕𝒆𝒇𝒆𝒖𝒊𝒍𝒍𝒆

é𝒒𝒖𝒊𝒑𝒐𝒏𝒅é𝒓é 𝒆𝒔𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕 𝒅𝒖 𝒓𝒊𝒔𝒒𝒖𝒆

Exercice 3 - Partie 1 : (5 points : 2+1+1+1)

ÉNONCÉ
Partie 1 :
𝑶𝒏 𝒄𝒐𝒏𝒔𝒊𝒅è𝒓𝒆 𝒅𝒆𝒖𝒙 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝑿𝟏 𝒆𝒕 𝑿𝟐 𝒏𝒐𝒓𝒎𝒂𝒍𝒆𝒔 𝒄𝒆𝒏𝒕𝒓é𝒆𝒔 ∶ 𝑬(𝑿𝒊 ) = 𝟎 , 𝒆𝒕
𝒓é𝒅𝒖𝒊𝒕𝒆𝒔 𝑽𝒂𝒓(𝑿𝒊 ) = 𝟏 ; 𝒊 = 𝟏, 𝟐 𝒂𝒚𝒂𝒏𝒕 𝒎ê𝒎𝒆 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 é𝒈𝒂𝒍𝒆 à 𝒄 ∶ 𝑪𝒐𝒗(𝑿𝟏 , 𝑿𝟐 ) = 𝒄 .
𝑶𝒏 𝒑𝒐𝒔𝒆 𝒀𝟏 = 𝑿𝟏 + 𝑿𝟐 𝒆𝒕 𝒀𝟐 = 𝑿𝟏 + 𝟐𝑿𝟐
1) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒆𝒔 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆𝒔 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆𝒔 𝒆𝒕 𝒍𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆𝒔 𝒅𝒆 𝒀𝟏 𝒆𝒕 𝒅𝒆 𝒀𝟐

315 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒀
2) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒎𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒅𝒖 𝒗𝒆𝒄𝒕𝒆𝒖𝒓 𝒀 = [ 𝟏 ]
𝒀𝟐
3) 𝑷𝒐𝒖𝒓 𝒒𝒖𝒆𝒍𝒍𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓 𝒅𝒆 𝒄 𝒍𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒀𝟏 𝒆𝒕 𝒀𝟐 𝒔𝒐𝒏𝒕 𝒆𝒍𝒍𝒆𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 ?

𝑱𝒖𝒔𝒕𝒊𝒇𝒊𝒆𝒓 𝒗𝒐𝒕𝒓𝒆 𝒓é𝒑𝒐𝒏𝒔𝒆.

4) 𝑷𝒐𝒖𝒓 𝒄𝒆𝒕𝒕𝒆 𝒗𝒂𝒍𝒆𝒖𝒓 𝒅𝒆 𝒄, 𝒄𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒅𝒆 𝒀𝟏 . 𝑷𝒓𝒐𝒖𝒗𝒆𝒓 𝒒𝒖𝒆 𝒄𝒆𝒕𝒕𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒆𝒔𝒕

𝒄𝒆𝒓𝒕𝒂𝒊𝒏𝒆 .

Corrigé
Partie 1 :
1)

∙ 𝑬(𝒀𝟏 ) = 𝑬(𝑿𝟏 + 𝑿𝟐 ) = 𝑬(𝑿

⏟ 𝟏 ) + 𝑬(𝑿
⏟ 𝟐 ) ⇒ 𝑬(𝒀𝟏 ) = 𝟎
𝟎 𝟎

∙ 𝑬(𝒀𝟏 ) = 𝑬(𝑿𝟏 + 𝟐𝑿𝟐 ) = 𝑬(𝑿

⏟ 𝟏 ) + 𝟐 𝑬(𝑿
⏟ 𝟐 ) ⇒ 𝑬(𝒀𝟐 ) = 𝟎
𝟎 𝟎

∙ 𝑽𝒂𝒓(𝒀𝟏 ) = 𝑽𝒂𝒓(𝑿𝟏 + 𝑿𝟐 ) = ⏟
𝑽𝒂𝒓(𝑿𝟏 ) + ⏟
𝑽𝒂𝒓(𝑿𝟐 ) + 𝟐 ⏟
𝑪𝒐𝒗(𝑿𝟏 , 𝑿𝟐 ) ⇒ 𝑽𝒂𝒓(𝒀𝟏 ) = 𝟐(𝟏 + 𝒄)
𝟏 𝟏 𝒄

𝑽𝒂𝒓(𝑿𝟏 ) + 𝟐𝟐 ⏟
∙ 𝑽𝒂𝒓(𝒀𝟐 ) = 𝑽𝒂𝒓(𝑿𝟏 + 𝟐𝑿𝟐 ) = ⏟ 𝑽𝒂𝒓(𝑿𝟐 ) + [𝟐 × 𝟏 × 𝟐 ⏟
𝑪𝒐𝒗(𝑿𝟏 , 𝑿𝟐 )]
𝟏 𝟏 𝒄

⇒ 𝑽𝒂𝒓(𝒀𝟐 ) = 𝟓 + 𝟒𝒄

2) 𝑶𝒏 𝒂 𝑽𝒂𝒓(𝒀𝟏 ) = 𝟐(𝟏 + 𝒄) 𝒆𝒕 𝑽𝒂𝒓(𝒀𝟐 ) = 𝟓 + 𝟒𝒄 . 𝑪𝒂𝒍𝒄𝒖𝒍𝒐𝒏𝒔 𝑪𝒐𝒗(𝒀𝟏 , 𝒀𝟐 ) ∶

𝑪𝒐𝒗(𝒀𝟏 , 𝒀𝟐 ) = 𝑪𝒐𝒗((𝑿𝟏 + 𝑿𝟐 ), (𝑿𝟏 + 𝟐𝑿𝟐 )) = 𝑪𝒐𝒗(𝑿𝟏 , (𝑿𝟏 + 𝟐𝑿𝟐 )) + 𝑪𝒐𝒗(𝑿𝟐 , (𝑿𝟏 + 𝟐𝑿𝟐 ))

= [𝑪𝒐𝒗(𝑿
⏟ 𝟏 , 𝑿𝟏 ) + 𝑪𝒐𝒗(𝑿𝟏 , 𝟐𝑿𝟐 )] + [𝑪𝒐𝒗(𝑿
⏟ 𝟐 , 𝑿𝟏 ) + 𝑪𝒐𝒗(𝑿𝟐 , 𝟐𝑿𝟐 )]
𝑽𝒂𝒓(𝑿𝟏 ) 𝒄

= [𝑽𝒂𝒓(𝑿𝟏 ) + 𝟐𝑪𝒐𝒗(𝑿𝟏 , 𝑿𝟐 )] + [𝒄 + 𝟐 𝑪𝒐𝒗(𝑿

⏟ 𝟐 , 𝑿𝟐 )] = [𝟏 + 𝟐𝒄] + [𝒄 + 𝟐]
𝑽𝒂𝒓(𝑿𝟐 )

𝑪𝒐𝒗(𝒀𝟏 , 𝒀𝟐 ) = 𝟑(𝟏 + 𝒄)

𝒀 𝟐(𝟏 + 𝒄) 𝟑(𝟏 + 𝒄)
𝑫′ 𝒐ù ∶ 𝛀𝒀 = 𝑽𝒂𝒓 ([ 𝟏 ]) = ( )
𝒀𝟐 𝟑(𝟏 + 𝒄) 𝟓 + 𝟒𝒄

𝑿𝒊 ↝ 𝓝(𝟎, 𝟏) ; 𝒊 = 𝟏, 𝟐
3)  𝑹𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒔 ∶ 𝑶𝒏 𝒂 ∶ {
𝒀𝟏 (𝒓𝒆𝒔𝒑. 𝒀𝟐 ) 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒄𝒐𝒎𝒃𝒊𝒏𝒂𝒊𝒔𝒐𝒏 𝒍𝒊𝒏é𝒂𝒊𝒓𝒆 𝒅𝒆 𝑿𝟏 𝒆𝒕 𝑿𝟐

𝑷𝒐𝒖𝒓 𝒒𝒖𝒆 𝒀𝟏 𝒆𝒕 𝒀𝟐 𝒔𝒐𝒊𝒆𝒏𝒕 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖é𝒆𝒔 𝒏𝒐𝒓𝒎𝒂𝒍𝒎𝒆𝒏𝒕 , 𝒊𝒍 𝒔𝒖𝒇𝒇𝒊𝒕 𝒅𝒐𝒏𝒄 𝒅′ 𝒂𝒗𝒐𝒊𝒓 𝑪𝒐𝒗(𝑿𝟏 , 𝑿𝟐 ) = 𝟎 ,

𝒂𝒖𝒕𝒓𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒕 𝒄 = 𝟎 𝒔𝒊 𝒏𝒐𝒏 𝒍𝒂 𝒔𝒕𝒂𝒃𝒊𝒍𝒊𝒕é 𝒏𝒆 𝒔𝒆𝒓𝒂 𝒑𝒂𝒔 𝒈𝒂𝒓𝒂𝒏𝒕𝒊𝒆 𝒆𝒕 𝒐𝒏 𝒏’𝒂 𝒑𝒂𝒔 𝒏é𝒄𝒆𝒔𝒔𝒂𝒊𝒓𝒆𝒎𝒆𝒏𝒕
𝒅𝒆𝒔 𝒗. 𝒂. 𝒏𝒐𝒓𝒎𝒂𝒍𝒆𝒔 ‼

316 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝑬𝒏 𝒆𝒇𝒇𝒆𝒕 𝒅𝒊𝒓𝒆 𝒒𝒖𝒆 𝒀𝟏 𝒆𝒕 𝒀𝟐 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊
𝑪𝒐𝒗(𝒀𝟏 , 𝒀𝟐 ) = 𝟎 (𝒆𝒏 𝒅′ 𝒂𝒖𝒕𝒓𝒆𝒔 𝒕𝒆𝒓𝒎𝒆𝒔 𝒄 = −𝟏)𝒄𝒆𝒍𝒂 𝒔𝒖𝒑𝒑𝒐𝒔𝒆 𝒒𝒖𝒆 𝒀𝟏 𝒆𝒕 𝒀𝟐 𝒔𝒐𝒊𝒆𝒏𝒕 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖é𝒆𝒔
𝒏𝒐𝒓𝒎𝒂𝒍𝒎𝒆𝒏𝒕 (𝒄 = 𝟎) 𝒅′ 𝒐ù 𝒍′ 𝒂𝒃𝒔𝒖𝒓𝒅𝒊𝒕é 𝒅𝒆 𝒍𝒂 𝒒𝒖𝒆𝒔𝒕𝒊𝒐𝒏 𝟑)
𝑹𝒆𝒇𝒐𝒓𝒎𝒖𝒍𝒐𝒏𝒔 𝒅𝒐𝒏𝒄 𝒍𝒂 𝒒𝒖𝒆𝒔𝒕𝒊𝒐𝒏 𝟑) ∶ 𝑷𝒐𝒖𝒓 𝒒𝒖𝒆𝒍𝒍𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓 𝒅𝒆 𝒄 𝒍𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒀𝟏 𝒆𝒕 𝒀𝟐 𝒔𝒐𝒏𝒕
𝒆𝒍𝒍𝒆𝒔 𝒏𝒐𝒏 𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔 ? 𝑱𝒖𝒔𝒕𝒊𝒇𝒊𝒆𝒓 𝒗𝒐𝒕𝒓𝒆 𝒓é𝒑𝒐𝒏𝒔𝒆.
𝒀𝟏 𝒆𝒕 𝒀𝟐 𝒔𝒐𝒏𝒕 𝒆𝒍𝒍𝒆𝒔 𝒏𝒐𝒏 𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔 ⇔ 𝑪𝒐𝒗(𝒀𝟏 , 𝒀𝟐 ) = 𝟎 ⇔ 𝟑(𝟏 + 𝒄) = 𝟎

𝒀𝟏 𝒆𝒕 𝒀𝟐 𝒔𝒐𝒏𝒕 𝒆𝒍𝒍𝒆𝒔 𝒏𝒐𝒏 𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔 ⇔ 𝒄 = −𝟏

4) 𝑨𝒊𝒏𝒔𝒊 𝒑𝒐𝒖𝒓 𝒄 = −𝟏 , 𝒐𝒏 𝒐𝒃𝒕𝒊𝒆𝒏𝒕 𝑽𝒂𝒓(𝒀𝟏 ) = 𝟐(𝟏 + (−𝟏)) = 𝟎

𝒆𝒕 𝒀𝟏 = 𝑬(𝒀𝟏 ) = 𝟎 (𝒑. 𝒔. ) ⇔ 𝒀𝟏 𝒆𝒔𝒕 𝒑𝒓𝒆𝒔𝒒𝒖𝒆 𝒔û𝒓𝒆𝒎𝒆𝒏𝒕 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆.

𝑷(𝒀𝟏 = 𝟎) = 𝟏
𝑬𝒏 𝒆𝒇𝒇𝒆𝒕 ∶ { 𝒅′ 𝒐ù 𝒀𝟏 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒄𝒆𝒓𝒕𝒂𝒊𝒏𝒆 𝒆𝒕 𝒐𝒏 𝒂𝒊𝒕
𝑷(𝒀𝟏 = 𝟎) = 𝟎 , 𝒑𝒐𝒖𝒓 𝒕𝒐𝒖𝒕 𝒚𝒊 ≠ 𝟎
𝒑𝒓𝒆𝒔𝒒𝒖𝒆 𝒔û𝒓𝒆 𝒒𝒖𝒆 𝒀𝟏 = 𝟎 .

𝑳𝒂 𝒍𝒐𝒊 𝒅𝒆 𝒀𝟏 𝒔𝒆𝒓𝒂 𝒅𝒐𝒏𝒄 𝒄𝒆𝒍𝒍𝒆 𝒅𝒆 𝒅𝒆 𝑫𝒊𝒓𝒂𝒄 𝜹𝟎 ∶ 𝒀𝟏 ↝ 𝜹𝟎

Exercice 4 : (10 points : 1 point par question)

ÉNONCÉ
𝑪𝒐𝒏𝒔𝒊𝒅é𝒓𝒐𝒏𝒔 𝒏 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒚𝟏 , 𝒚𝟐 , … , 𝒚𝒏 𝒂𝒚𝒂𝒏𝒕 𝒑𝒐𝒖𝒓 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆
𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 𝒎 𝒆𝒕 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝝈𝟐 :
𝒏 𝒏
𝑬(𝒚𝒊 ) = 𝒎 𝟏 𝟐
𝟏
{ 𝒑𝒐𝒖𝒓 𝒊 = 𝟏, 𝟐, … , 𝒏 ; 𝒐𝒏 𝒏𝒐𝒕𝒆 ̅
𝒚 𝒏 = ∑ 𝒚 𝒊 𝒆𝒕 𝑺 = ̅ 𝒏 )𝟐
∑(𝒚𝒊 − 𝒚
𝑽(𝒚𝒊 ) = 𝝈𝟐 𝒏 𝒏
𝒊=𝟏 𝒊=𝟏

̅𝒏 ) 𝒆𝒕 𝑽(𝒚
1) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝑬(𝒚 ̅𝒏 )
2)
a) 𝑬𝒙𝒑𝒓𝒊𝒎𝒆𝒓 𝒚 ̅𝒏 𝒆𝒏 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒚
̅𝒏−𝟏
b) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒆𝒏𝒕𝒓𝒆 𝒚 ̅𝒏−𝟏 𝒆𝒕 𝒚
̅𝒏
̅𝒏−𝟏 𝒆𝒕 𝒚
c) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅𝒆 𝒄𝒐𝒓𝒓é𝒍𝒂𝒕𝒊𝒐𝒏 𝒍𝒊𝒏é𝒂𝒊𝒓𝒆 𝒆𝒏𝒕𝒓𝒆 𝒚 ̅𝒏
d) 𝑰𝒏𝒕𝒆𝒓𝒑𝒓é𝒕𝒆𝒓 𝒄𝒆 𝒅𝒆𝒓𝒏𝒊𝒆𝒓 𝒓é𝒔𝒖𝒍𝒕𝒂𝒕

317 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
3)
a) 𝑻𝒓𝒐𝒖𝒗𝒆𝒓 𝒖𝒏 𝒎𝒊𝒏𝒐𝒓𝒂𝒏𝒕 𝒑𝒐𝒖𝒓 𝑷[−𝜶 ≤ 𝒚 ̅𝒏 − 𝒎 ≤ 𝜶], 𝒑𝒐𝒖𝒓 𝒖𝒏 𝒔𝒄𝒂𝒍𝒂𝒊𝒓𝒆 𝒑𝒐𝒔𝒊𝒕𝒊𝒇 𝜶
̅𝒏 𝒒𝒖𝒂𝒏𝒅 𝒏 → +∞
b) 𝑬𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒍𝒂 𝒍𝒊𝒎𝒊𝒕𝒆 𝒆𝒏 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒅𝒆 𝒚
4)
a) 𝑷𝒓𝒐𝒖𝒗𝒆𝒓 𝒒𝒖𝒆 𝒍’𝒐𝒏 𝒂 :
𝒏

𝒏𝑺 = ∑(𝒚𝒊 − 𝒎)𝟐 − 𝒏(𝒚

𝟐
̅𝒏 − 𝒎)𝟐
𝒊=𝟏
𝒏𝑺𝟐
b) 𝑬𝒏 𝒂𝒅𝒎𝒆𝒕𝒕𝒂𝒏𝒕 𝒒𝒖𝒆 𝒍𝒆𝒔 𝒚𝒊 𝒔𝒐𝒏𝒕 𝒅𝒆𝒔 𝒍𝒐𝒊𝒔 𝒏𝒐𝒓𝒎𝒂𝒍𝒆𝒔, 𝒑𝒓𝒐𝒖𝒗𝒆𝒓 𝒒𝒖𝒆 𝒆𝒔𝒕 𝒍𝒂
𝝈𝟐

𝒅𝒊𝒇𝒇é𝒓𝒆𝒏𝒄𝒆 𝒅𝒆 𝒅𝒆𝒖𝒙 𝒍𝒐𝒊𝒔 𝒅𝒆 𝑲𝒉𝒊‑𝑫𝒆𝒖𝒙 𝒅𝒐𝒏𝒕 𝒊𝒍 𝒇𝒂𝒖𝒕 𝒑𝒓é𝒄𝒊𝒔𝒆𝒓 𝒍𝒆𝒔 𝒅𝒆𝒈𝒓é𝒔 𝒅𝒆 𝒍𝒊𝒃𝒆𝒓𝒕é
c) 𝑬𝒏 𝒂𝒅𝒎𝒆𝒕𝒕𝒂𝒏𝒕 𝒒𝒖𝒆 𝒚 ̅𝒏 𝒆𝒕 ∑𝒏𝒊=𝟏(𝒚𝒊 − 𝒎)𝟐 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔,
𝒏𝑺𝟐
𝒆𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝟐
𝝈
Corrigé
1)
𝒏 𝒏 𝒏 𝒏
𝟏 𝟏 𝟏 𝟏 𝟏
̅𝒏 ) = 𝑬 ( ∑ 𝒚𝒊 ) = 𝑬 (∑ 𝒚𝒊 ) = ∑ 𝑬(𝒚𝒊 ) = ∑ 𝒎 = (𝒏. 𝒎). 𝑫′ 𝒐ù 𝑬(𝒚
∙ 𝑬(𝒚 ̅𝒏 ) = 𝑬(𝒚𝒊 ) = 𝒎
𝒏 𝒏 𝒏 𝒏 𝒏
𝒊=𝟏 𝒊=𝟏 𝒊=𝟏 𝒊=𝟏
𝒏 𝒏
𝟏 𝟏
̅𝒏 ) = 𝑽 ( ∑ 𝒚𝒊 ) = 𝟐 𝑽 (∑ 𝒚𝒊 )
∙ 𝑽(𝒚
𝒏 𝒏
𝒊=𝟏 𝒊=𝟏
𝒏 𝒏

𝒐𝒓 𝒚𝟏 , 𝒚𝟐 , … , 𝒚𝒏 𝒔𝒐𝒏𝒕 𝒅𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 ⇒ 𝑽 (∑ 𝒚𝒊 ) = ∑ 𝑽(𝒚𝒊 )

𝒊=𝟏 𝒊=𝟏
𝒏 𝒏 𝒏
𝟏 𝟏 𝟏 𝟐
𝟏 𝟐)
𝝈𝟐
̅𝒏 ) =
𝒑𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 𝑽(𝒚 𝑽 (∑ 𝒚 𝒊 ) = ∑ 𝑽(𝒚 𝒊 ) = ∑ 𝝈 = (𝒏. 𝝈 =
𝒏𝟐 𝒏𝟐 𝒏𝟐 𝒏 𝒏
𝒊=𝟏 𝒊=𝟏 𝒊=𝟏

𝑽(𝒚𝒊 ) 𝝈𝟐
̅𝒏 ) =
𝑽(𝒚 =
𝒏 𝒏
2)
a)
𝒏 𝒏−𝟏
𝟏 𝟏 𝟏 𝒏−𝟏 𝟏
̅𝒏 = ∑ 𝒚𝒊 = (𝒚𝒏 + ∑ 𝒚𝒊 ) = [𝒚𝒏 + (𝒏 − 𝟏)𝒚
𝒚 ̅𝒏−𝟏 ] ⇒ 𝒚
̅𝒏 = ( )𝒚
̅𝒏−𝟏 + 𝒚𝒏
𝒏 𝒏 𝒏 𝒏 𝒏
𝒊=𝟏 𝒊=𝟏

b)

𝒏−𝟏 𝟏 𝒏−𝟏 𝟏
𝑪𝒐𝒗(𝒚 ̅𝒏 ) = 𝑪𝒐𝒗 (𝒚
̅𝒏−𝟏 , 𝒚 ̅𝒏−𝟏 , ( )𝒚
̅𝒏−𝟏 + 𝒚𝒏 ) = ̅𝒏−𝟏 ) + 𝑪𝒐𝒗(𝒚
̅𝒏−𝟏 , 𝒚
𝑪𝒐𝒗(𝒚 ̅𝒏−𝟏 , 𝒚𝒏 )
𝒏 𝒏 𝒏 𝒏
𝒏−𝟏 𝟏
̅𝒏 ) =
̅𝒏−𝟏 , 𝒚
𝑪𝒐𝒗(𝒚 ̅𝒏−𝟏 ) + 𝑪𝒐𝒗(𝒚
𝑽(𝒚 ̅𝒏−𝟏 , 𝒚𝒏 )
𝒏 𝒏

318 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒐𝒓 𝒍𝒆𝒔 𝒚𝟏 , 𝒚𝟐 , … , 𝒚𝒏 𝒔𝒐𝒏𝒕 𝒅𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔
𝒏−𝟏 𝒏−𝟏 𝒏−𝟏
𝟏 𝟏 𝟏
̅𝒏−𝟏 , 𝒚𝒏 ) = 𝑪𝒐𝒗 (
𝑪𝒐𝒗(𝒚 ∑ 𝒚𝒊 , 𝒚𝒏 ) = 𝑪𝒐𝒗 (∑ 𝒚𝒊 , 𝒚𝒏 ) = 𝑪𝒐𝒗(𝒚𝒊 , 𝒚𝒏 )
∑⏟
𝒏−𝟏 𝒏−𝟏 𝒏−𝟏
𝒊=𝟏 𝒊=𝟏 𝒊=𝟏 𝟎 ; ∀𝒊≠𝒏

̅𝒏−𝟏 , 𝒚𝒏 ) = 𝟎
⇒ 𝑪𝒐𝒗(𝒚

′
𝝈𝟐 𝒏−𝟏 𝝈𝟐
̅𝒏−𝟏 ) =
𝑫 𝒂𝒖𝒕𝒓𝒆 𝒑𝒂𝒓𝒕 𝑽(𝒚 , 𝒂𝒊𝒏𝒔𝒊 𝑪𝒐𝒗(𝒚 ̅𝒏 ) = (
̅𝒏−𝟏 , 𝒚 )( )
𝒏−𝟏 𝒏 𝒏−𝟏
𝝈𝟐
̅𝒏 ) = 𝑽(𝒚
̅𝒏−𝟏 , 𝒚
𝑫’𝒐ù 𝑪𝒐𝒗(𝒚 ̅𝒏 ) =
𝒏

̅𝒏 )
̅𝒏−𝟏 ,𝒚
𝑪𝒐𝒗(𝒚 𝝈𝟐 ⁄𝒏 𝝈𝟐 √(𝒏−𝟏)𝒏 𝟏
c) 𝝆𝒚̅𝒏−𝟏 ,𝒚̅𝒏 = = = (𝒏)( ) ⇒ 𝝆𝒚̅𝒏−𝟏 ,𝒚̅𝒏 = √𝟏 −
̅𝒏−𝟏 )𝑽(𝒚
√𝑽(𝒚 ̅𝒏 ) 𝝈𝟐 𝒏
𝝈𝟐 𝝈𝟐
√( )( )
𝒏−𝟏 𝒏

d) 𝑷𝒐𝒖𝒓 𝒏 𝒂𝒔𝒔𝒆𝒛 𝒈𝒓𝒂𝒏𝒅, 𝒐𝒏 𝒐𝒃𝒕𝒊𝒆𝒏𝒕 𝒖𝒏𝒆 𝒄𝒐𝒓𝒓é𝒍𝒂𝒕𝒊𝒐𝒏 𝒑𝒐𝒔𝒊𝒕𝒊𝒗𝒆 𝒑𝒂𝒓𝒇𝒂𝒊𝒕𝒆 :

𝝆𝒚̅𝒏−𝟏 ,𝒚̅𝒏 > 𝟎 ⇒𝒚̅𝒏−𝟏 𝒆𝒕 𝒚

̅𝒏 𝒗𝒂𝒓𝒊𝒆𝒏𝒕 𝒅𝒂𝒏𝒔 𝒍𝒆 𝒎ê𝒎𝒆 𝒔𝒆𝒏𝒔
{𝒍𝒊𝒎 𝝆 =𝟏 ⇒ 𝒑𝒐𝒖𝒓 𝒏 𝒂𝒔𝒔𝒆𝒛 𝒈𝒓𝒂𝒏𝒅, 𝒚̅𝒏−𝟏 𝒆𝒕 𝒚
̅𝒏 𝒐𝒏𝒕 𝒅𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 𝒗𝒐𝒊𝒔𝒊𝒏𝒆𝒔
̅𝒏−𝟏 ,𝒚
𝒚 ̅𝒏
𝒏→∞

3)
̅𝒏 ) = 𝒎
̅𝒏 − 𝒎 ≤ 𝜶], 𝒐𝒓 𝑬(𝒚
a) 𝑶𝒏 𝒂 ∶ 𝑷[−𝜶 ≤ 𝒚

̅𝒏 − 𝒎 ≤ 𝜶] = 𝑷[−𝜶 ≤ 𝒚
𝒅𝒐𝒏𝒄 𝑷[−𝜶 ≤ 𝒚 ̅𝒏 ) ≤ 𝜶] = 𝑷[|𝒚
̅𝒏 − 𝑬(𝒚 ̅𝒏 )| ≤ 𝜶 ]
̅𝒏 − 𝑬(𝒚
̅𝒏 )| > 𝜶 ]
̅𝒏 − 𝑬(𝒚
= 𝟏 − 𝑷[|𝒚
̅𝒏 )
𝑽(𝒚
̅𝒏 )| > 𝜶 ] ≤
̅𝒏 − 𝑬(𝒚
𝑫’𝒂𝒑𝒓è𝒔 𝒍’𝒊𝒏é𝒈𝒂𝒍𝒊𝒕é 𝒅𝒆 𝑩𝒊𝒆𝒏𝒂𝒚𝒎é 𝑻𝒄𝒉𝒆𝒃𝒚𝒄𝒉𝒆𝒗 𝒐𝒏 𝒂 𝑷[|𝒚
𝜶𝟐
𝝈𝟐 𝝈𝟐 𝝈𝟐
̅𝒏 ) =
𝑽(𝒚 𝒑𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 𝑷[|𝒚 ̅𝒏 )| > 𝜶 ] ≤
̅𝒏 − 𝑬(𝒚 ⇔ 𝟏 − ̅
𝑷[|𝒚 𝒏 − ̅
𝑬(𝒚 𝒏 )| > 𝜶 ] ≥ 𝟏 −
𝒏 𝒏𝜶𝟐 𝒏𝜶𝟐
𝝈𝟐 𝝈𝟐
𝑫′ 𝒐ù 𝑷[−𝜶 ≤ 𝒚
̅𝒏 − 𝒎 ≤ 𝜶] ≥ 𝟏 − , 𝟏 − ̅𝒏 − 𝒎 ≤ 𝜶]
é𝒕𝒂𝒏𝒕 𝒖𝒏 𝒎𝒊𝒏𝒐𝒓𝒂𝒏𝒕 𝒑𝒐𝒖𝒓 𝑷[−𝜶 ≤ 𝒚
𝒏𝜶𝟐 𝒏𝜶𝟐
𝝈𝟐
̅𝒏 − 𝒎 ≤ 𝜶] ≤ 𝟏
b) 𝑶𝒏 𝒂 𝒂𝒊𝒏𝒔𝒊 ∶ 𝟏 − 𝒏𝜶𝟐 ≤ 𝑷[−𝜶 ≤ 𝒚

𝝈𝟐
𝑶𝒓 𝐥𝐢𝐦 (𝟏 − ̅𝒏 − 𝒎 ≤ 𝜶] = 𝟏 𝒐𝒖 𝒆𝒏𝒄𝒐𝒓𝒆 𝐥𝐢𝐦 𝑷[|𝒚
) = 𝟏 ⇒ 𝐥𝐢𝐦 𝑷[−𝜶 ≤ 𝒚 ̅𝒏 − 𝒎| ≤ 𝜶] = 𝟏
𝒏→∞ 𝒏𝜶𝟐 𝒏→∞ 𝒏→∞

𝑷
𝑫′ 𝒐ù 𝒚
̅𝒏 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒗𝒆𝒓𝒔 𝒎 ∶ 𝒚
̅𝒏 ⟶ 𝒎

4)
a) 𝒏𝑺𝟐 = ∑𝒏𝒊=𝟏(𝒚𝒊 − 𝒚
̅𝒏 )𝟐 = ∑𝒏𝒊=𝟏[(𝒚𝒊 − 𝒎) − (𝒚
̅𝒏 − 𝒎)]𝟐
𝒏

𝒏𝑺𝟐 = ∑[(𝒚𝒊 − 𝒎)𝟐 − 𝟐(𝒚 ̅𝒏 − 𝒎)𝟐 ]

̅𝒏 − 𝒎)(𝒚𝒊 − 𝒎) + (𝒚
𝒊=𝟏

319 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒏 𝒏 𝒏
𝟐
̅𝒏 − 𝒎)𝟐
̅𝒏 − 𝒎) ∑(𝒚𝒊 − 𝒎) + ∑(𝒚
= ∑(𝒚𝒊 − 𝒎) − 𝟐(𝒚
𝒊=𝟏 𝒊=𝟏 𝒊=𝟏

𝒏 𝒏 𝒏
𝟐
̅𝒏 − 𝒎)𝟐
̅𝒏 − 𝒎) ∑ 𝒚𝒊 − ∑ 𝒎 + 𝒏(𝒚
= ∑(𝒚𝒊 − 𝒎) − 𝟐(𝒚
𝒊=𝟏 ⏟
𝒊=𝟏 ⏟
𝒊=𝟏
( ̅
𝒏𝒚 𝒏 𝒏.𝒎 )

= ∑(𝒚𝒊 − 𝒎)𝟐 − 𝟐(𝒚

̅𝒏 − 𝒎)(𝒏(𝒚 ̅𝒏 − 𝒎)𝟐
̅𝒏 − 𝒎)) + 𝒏(𝒚
𝒊=𝟏
𝒏

= ∑(𝒚𝒊 − 𝒎)𝟐 − 𝟐𝒏(𝒚

̅𝒏 − 𝒎)𝟐 + 𝒏(𝒚
̅𝒏 − 𝒎)𝟐
𝒊=𝟏

𝑫′ 𝒐ù 𝒏𝑺𝟐 = ∑(𝒚𝒊 − 𝒎)𝟐 − 𝒏(𝒚

̅𝒏 − 𝒎)𝟐
𝒊=𝟏

b)
𝒏
𝟐)
𝒚𝒊 − 𝒎 (𝒚𝒊 − 𝒎)𝟐 𝟐 (𝟏)
(𝒚𝒊 − 𝒎)𝟐
𝒚𝒊 ↝ 𝓝(𝒎, 𝝈 ⇒ ↝ 𝓝(𝟎, 𝟏) ⇒ ↝ 𝝌 ⇒ 𝑯 = ∑ ↝ 𝝌𝟐 (𝒏) (𝟏)
𝝈 𝝈𝟐 𝝈𝟐
𝒊=𝟏

𝑫′ 𝒂𝒖𝒕𝒓𝒆 𝒑𝒂𝒓𝒕 ∀ 𝒊 ∈ {𝟏, 𝟐, … , 𝒏} 𝒚𝒊 ↝ 𝓝(𝒎, 𝝈𝟐 ) 𝒐𝒏 𝒂 𝒚 ̅𝒏 ), 𝑽(𝒚

̅𝒏 ↝ 𝓝(𝑬(𝒚 ̅𝒏 ))

𝑽(𝒚𝒊 ) 𝝈𝟐 𝝈 𝟐
̅𝒏 ) = 𝑬(𝒚𝒊 ) = 𝒎 𝒆𝒕 𝑽(𝒚
𝒐𝒓 𝑬(𝒚 ̅𝒏 ) = = =( )
𝒏 𝒏 √𝒏
𝝈 𝟐 ̅𝒏 − 𝒎
𝒚 ̅𝒏 − 𝒎
𝒚
̅𝒏 ↝ 𝓝 (𝒎, (
𝒐𝒏 𝒐𝒃𝒕𝒊𝒆𝒏𝒕 𝒚 ) )⇒ ↝ 𝓝(𝟎, 𝟏) ⇒ ( ) ↝ 𝝌𝟐 (𝟏)
√𝒏 𝝈⁄√𝒏 𝝈⁄√𝒏
̅𝒏 − 𝒎)𝟐
𝒏(𝒚
⇒𝑼= 𝟐
↝ 𝝌𝟐 (𝟏) (𝟐)
𝝈
𝒏
𝒏𝑺𝟐 ∑𝒏𝒊=𝟏(𝒚𝒊 − 𝒎)𝟐 − 𝒏(𝒚
̅𝒏 − 𝒎)𝟐 (𝒚𝒊 − 𝒎)𝟐 𝒏(𝒚
̅𝒏 − 𝒎)𝟐
𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 𝟐 = =∑ − =𝑯−𝑼
𝝈 𝝈𝟐 𝝈𝟐 𝝈𝟐
𝒊=𝟏

𝒏𝑺𝟐
𝑫 𝒐ù 𝑾 = 𝟐 = 𝑯 − 𝑼 𝒂𝒗𝒆𝒄 𝑯 ↝ 𝝌𝟐 (𝒏) 𝒆𝒕 𝑼 ↝ 𝝌𝟐 (𝟏)
′
𝝈
̅𝒏 𝒆𝒕 ∑𝒏𝒊=𝟏(𝒚𝒊 − 𝒚
c) 𝑹𝒆𝒇𝒐𝒓𝒎𝒖𝒍𝒐𝒏𝒔 𝒅𝒐𝒏𝒄 𝒍𝒂 𝒒𝒖𝒆𝒔𝒕𝒊𝒐𝒏 𝟒) 𝐛) ∶ 𝑬𝒏 𝒂𝒅𝒎𝒆𝒕𝒕𝒂𝒏𝒕 𝒒𝒖𝒆 𝒚 ̅ 𝒏 )𝟐
𝒏𝑺𝟐
𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔, 𝒆𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝟐
𝝈
𝒏

̅𝒏 )𝟐 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔, 𝒅’𝒂𝒑𝒓è𝒔 𝒍’𝒉𝒚𝒑𝒐𝒕𝒉è𝒔𝒆 𝒅𝒆 𝒍’𝒆𝒙𝒆𝒓𝒄𝒊𝒄𝒆.

̅𝒏 𝒆𝒕 ∑(𝒚𝒊 − 𝒚
𝒚
𝒊=𝟏

320 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝑿 𝒏(𝒀 − 𝒎)𝟐
𝑺𝒐𝒊𝒆𝒏𝒕, 𝒇 𝒆𝒕 𝒈 𝒅𝒆𝒖𝒙 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏𝒔 𝒕𝒆𝒍𝒍𝒆𝒔 𝒒𝒖𝒆 ∶ 𝒇(𝑿) = 𝟐 𝒆𝒕 𝒈(𝒀) = .
𝝈 𝝈𝟐
𝒏 𝒏
𝟏
𝑶𝒏 𝒗é𝒓𝒊𝒇𝒊𝒆 𝒇𝒂𝒄𝒊𝒍𝒆𝒎𝒆𝒏𝒕 𝒒𝒖𝒆 : 𝒇 (∑(𝒚𝒊 − 𝒚
̅𝒏 )𝟐 ̅ 𝒏 )𝟐 = 𝑾
) = 𝟐 ∑(𝒚𝒊 − 𝒚
𝝈
𝒊=𝟏 𝒊=𝟏

̅𝒏 − 𝒎)𝟐
𝒏(𝒚
̅𝒏 ) =
𝒆𝒕 𝒒𝒖𝒆 𝒈(𝒚 =𝑼
𝝈𝟐
𝑫’𝒂𝒖𝒕𝒓𝒆 𝒑𝒂𝒓𝒕 ∶ 𝒇 𝒆𝒕 𝒈 𝒅𝒆𝒖𝒙 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏𝒔 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆𝒔 𝒆𝒕 ∶
̅𝒏 )] = 𝑬(𝑼) = 𝟏 , 𝒆𝒙𝒊𝒔𝒕𝒆 𝒆𝒕 𝒇𝒊𝒏𝒊𝒆 , 𝒄𝒂𝒓 𝑼 ↝ 𝝌𝟐 (𝟏)
𝑬[𝒈(𝒚
𝒏

̅𝒏 )𝟐 )] = 𝑬(𝑾) = 𝑬(𝑯 − 𝑼) = 𝑬(𝑯) − 𝑬(𝑼) = 𝒏 − 𝟏, 𝒆𝒙𝒊𝒔𝒕𝒆 𝒆𝒕 𝒇𝒊𝒏𝒊𝒆 ,

𝒂𝒊𝒏𝒔𝒊 𝒒𝒖𝒆 𝑬 [ 𝒇 (∑(𝒚𝒊 − 𝒚
𝒊=𝟏

𝒄𝒂𝒓 𝑾 = 𝑯 − 𝑼 , 𝑬(𝑼) = 𝟏 𝒆𝒕 𝑯 ↝ 𝝌𝟐 (𝒏), 𝒅𝒐𝒏𝒄 𝑬(𝑯) = 𝒏

𝒏

̅𝒏 )𝟐 ) 𝒆𝒕 𝑼 = 𝒈(𝒚
𝑰𝒍 𝒆𝒏 𝒓é𝒔𝒖𝒍𝒕𝒆 𝒆𝒏𝒇𝒊𝒏 𝒒𝒖𝒆 ∶ 𝑾 = 𝒇 (∑(𝒚𝒊 − 𝒚 ̅𝒏 ) 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔.
𝒊=𝟏

𝑶𝒏 𝒔𝒆 𝒑𝒓𝒐𝒑𝒐𝒔𝒆 𝒎𝒂𝒊𝒏𝒕𝒆𝒏𝒂𝒏𝒕 𝒅𝒆 𝒅é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒅𝒆

𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝑾 = 𝑯 − 𝑼:
𝑶𝒓 𝑾 = 𝑯 − 𝑼, 𝒅𝒐𝒏𝒄 𝑯 = 𝑾 + 𝑼. 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒅’𝒂𝒃𝒐𝒓𝒅 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔
𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒅𝒆 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝑯 ∶ 𝑴𝑯 (𝒕) = 𝑴𝑾+𝑼 (𝒕) = 𝑴𝑾 (𝒕)𝑴𝑼 (𝒕) , 𝒄𝒂𝒓 𝒍𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔
𝑴𝑯 (𝒕)
𝑾 𝒆𝒕 𝑼 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 ⇒ 𝑴𝑾 (𝒕) =
𝑴𝑼 (𝒕)
𝒏 𝟏
𝑯 ↝ 𝝌𝟐 (𝒏) ⇒ 𝑴𝑯 (𝒕) = (𝟏 − 𝟐𝒕)−𝟐 , ∀ 𝒕 < 𝒏
𝟐 𝑴𝑯 (𝒕) (𝟏 − 𝟐𝒕)−𝟐
𝒐𝒓 𝒆𝒕 ⇒ 𝑴𝑾 (𝒕) = =
𝟏 𝟏 𝑴𝑼 (𝒕) (𝟏 − 𝟐𝒕)−𝟏𝟐
𝑼 ↝ 𝝌𝟐 (𝟏) ⇒ 𝑴𝑼 (𝒕) = (𝟏 − 𝟐𝒕)−𝟐 , ∀ 𝒕 <
{ 𝟐
(𝒏−𝟏) 𝒏𝑺𝟐
𝑰𝒍 𝒆𝒏 𝒓é𝒔𝒖𝒍𝒕𝒆 𝒒𝒖𝒆 𝑴𝑾 (𝒕) = (𝟏 − 𝟐𝒕)− 𝟐 ′
𝒅 𝒐ù 𝟐
↝ 𝝌𝟐 (𝒏 − 𝟏)
𝝈

321 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014

Exercice 5 : (10 points : 1 point par question)

ÉNONCÉ
𝑶𝒏 𝒄𝒐𝒏𝒔𝒊𝒅è𝒓𝒆 𝒖𝒏𝒆 𝒈𝒓𝒂𝒏𝒅𝒆𝒖𝒓 𝒙𝒕 𝒑𝒐𝒔𝒊𝒕𝒊𝒗𝒆 𝒕𝒆𝒍𝒍𝒆 𝒒𝒖𝒆 :
𝒚𝒕 = 𝒍𝒏 𝒙𝒕 = 𝒂𝒕 + 𝒃 + 𝒖𝒕 𝒑𝒐𝒖𝒓 𝒕 = 𝟏, 𝟐, … , 𝑻
𝑨𝒗𝒆𝒄 𝒖𝒕 𝒅𝒆𝒔 𝒕𝒆𝒓𝒎𝒆𝒔 𝒅’𝒆𝒓𝒓𝒆𝒖𝒓𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒔, 𝒔𝒖𝒊𝒗𝒂𝒏𝒕 𝒍𝒂 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝒄𝒆𝒏𝒕𝒓é𝒆 𝒓é𝒅𝒖𝒊𝒕𝒆 :
𝑬(𝒖𝒕 ) = 𝟎 ; 𝑽(𝒖𝒕 ) = 𝟏. 𝑶𝒏 𝒑𝒐𝒔𝒆 𝚫𝒚𝒕 = 𝒚𝒕 − 𝒚𝒕−𝟏
1) 𝑷𝒓𝒐𝒖𝒗𝒆𝒓 𝒒𝒖𝒆 𝒔𝒊 𝒐𝒏 𝒏é𝒈𝒍𝒊𝒈𝒆 𝒍𝒆 𝒕𝒆𝒓𝒎𝒆 𝒅’𝒆𝒓𝒓𝒆𝒖𝒓, 𝒍𝒆 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆 𝒂 𝒆𝒔𝒕 𝒍𝒆 𝒕𝒂𝒖𝒙 𝒅𝒆
𝒄𝒓𝒐𝒊𝒔𝒔𝒂𝒏𝒄𝒆 𝒎𝒐𝒚𝒆𝒏 𝒅𝒆 𝒍𝒂 𝒈𝒓𝒂𝒏𝒅𝒆𝒖𝒓 𝒚𝒕
2) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍’𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 𝒆𝒕 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒅𝒆 𝚫𝒚𝒕
3) 𝑷𝒓𝒐𝒖𝒗𝒆𝒓 𝒒𝒖𝒆 𝒍𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅𝒆 𝒄𝒐𝒓𝒓é𝒍𝒂𝒕𝒊𝒐𝒏 𝒍𝒊𝒏é𝒂𝒊𝒓𝒆 𝒆𝒏𝒕𝒓𝒆 𝚫𝒚𝒕 𝒆𝒕 𝚫𝒚𝒕−𝟏 𝒆𝒔𝒕
𝟏
é𝒈𝒂𝒍𝒆 à −
𝟐
4) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅𝒆 𝒄𝒐𝒓𝒓é𝒍𝒂𝒕𝒊𝒐𝒏 𝒍𝒊𝒏é𝒂𝒊𝒓𝒆 𝒆𝒏𝒕𝒓𝒆 𝚫𝒚𝒕 𝒆𝒕 𝚫𝒚𝒕−𝟐
5) 𝑶𝒏 𝒔𝒖𝒑𝒑𝒐𝒔𝒆 𝒅𝒂𝒏𝒔 𝒄𝒆𝒕𝒕𝒆 𝒒𝒖𝒆𝒔𝒕𝒊𝒐𝒏 𝒒𝒖𝒆 𝒃 = 𝟎
i. 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒆𝒏 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 𝒅𝒆 𝒚𝒕 𝒆𝒕 𝒅𝒆 𝒕 𝒑𝒐𝒖𝒓 𝒕 = 𝟏, 𝟐, … , 𝑻
𝒍’𝒆𝒙𝒑𝒓𝒆𝒔𝒔𝒊𝒐𝒏 𝒅𝒆 𝒍’𝒆𝒔𝒕𝒊𝒎𝒂𝒕𝒆𝒖𝒓 𝒅𝒆 𝒂 𝒑𝒂𝒓 𝒍𝒆𝒔 𝒎𝒐𝒊𝒏𝒅𝒓𝒆𝒔 𝒄𝒂𝒓𝒓é𝒔 𝒐𝒓𝒅𝒊𝒏𝒂𝒊𝒓𝒆𝒔
ii. 𝑬𝒙𝒑𝒓𝒊𝒎𝒆𝒓 𝒄𝒆𝒕𝒕𝒆 𝒆𝒔𝒕𝒊𝒎𝒂𝒕𝒊𝒐𝒏 𝒆𝒏 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 𝒅𝒆 𝒙𝒕 𝒆𝒕 𝒅𝒆 𝒕
𝒑𝒐𝒖𝒓 𝒕 = 𝟏, 𝟐, … , 𝑻
6) 𝑶𝒏 𝒔𝒖𝒑𝒑𝒐𝒔𝒆 𝒅𝒂𝒏𝒔 𝒄𝒆𝒕𝒕𝒆 𝒒𝒖𝒆𝒔𝒕𝒊𝒐𝒏 𝒒𝒖𝒆 𝒂 = 𝒃 = 𝟎
i. 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒙𝒕
ii. 𝑬𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒍𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒅𝒆 𝒙𝒕
iii. 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒈é𝒐𝒎é𝒕𝒓𝒊𝒒𝒖𝒆 𝒅𝒆 𝒙𝒕 𝒑𝒐𝒖𝒓 𝒕 = 𝟏, 𝟐, … , 𝑻 𝒆𝒏 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏
𝒅𝒆 𝒍𝒂 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒂𝒓𝒊𝒕𝒉𝒎é𝒕𝒊𝒒𝒖𝒆 𝒅𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒖𝒕
iv. 𝑬𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒍’𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 𝒅𝒆 𝒄𝒆𝒕𝒕𝒆 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒈é𝒐𝒎é𝒕𝒓𝒊𝒒𝒖𝒆

Corrigé
1) 𝚫𝒚𝒕 = 𝒚𝒕 − 𝒚𝒕−𝟏 = (𝒂𝒕 + 𝒃 + 𝒖𝒕 ) − (𝒂(𝒕 − 𝟏) + 𝒃 + 𝒖𝒕−𝟏 )
= (𝒂𝒕 + 𝒃 + 𝒖𝒕 ) − (𝒂𝒕 − 𝒂 + 𝒃 + 𝒖𝒕−𝟏 ) = 𝒂 + (𝒖𝒕 − 𝒖𝒕−𝟏 ) = 𝒂 + 𝚫𝒖𝒕

𝑬𝒏 𝒏é𝒈𝒍𝒊𝒈𝒆𝒂𝒏𝒕 𝒍𝒆𝒔 𝒑𝒆𝒓𝒕𝒖𝒓𝒃𝒂𝒕𝒊𝒐𝒏𝒔 (𝚫𝒖𝒕 ≅ 𝟎) 𝒐𝒏 𝒐𝒃𝒕𝒊𝒆𝒏𝒕 𝚫𝒚𝒕 ≅ 𝒂 , 𝒕𝒂𝒖𝒙 𝒅𝒆 𝒄𝒓𝒐𝒊𝒔𝒔𝒂𝒏𝒄𝒆

𝒚𝒕 − 𝒚𝒕−𝟏 𝜟𝒚𝒕
𝒎𝒐𝒚𝒆𝒏 𝒅𝒆 𝒍𝒂 𝒈𝒓𝒂𝒏𝒅𝒆𝒖𝒓 , (𝒚𝒕 ) ∶ = = 𝜟𝒚𝒕
𝒕 − (𝒕 − 𝟏) 𝟏
2)

322 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
∙ 𝑬(𝚫𝒚𝒕 ) = 𝑬(𝒂 + (𝒖𝒕 − 𝒖𝒕−𝟏 )) = 𝒂 + 𝑬(𝒖 𝑬(𝒖𝒕−𝟏 ) . 𝑫′ 𝒐ù 𝑬(𝚫𝒚𝒕 ) = 𝒂
⏟ 𝒕) − ⏟
𝟎 𝟎

∙ 𝑽(𝚫𝒚𝒕 ) = 𝑽(𝒂 + (𝒖𝒕 − 𝒖𝒕−𝟏 )) = 𝑽(𝒖𝒕 − 𝒖𝒕−𝟏 )

𝒐𝒓 𝒍𝒆𝒔 𝒕𝒆𝒓𝒎𝒆𝒔 𝒅’𝒆𝒓𝒓𝒆𝒖𝒓𝒔 (𝒖𝒕 )𝒇𝒐𝒓𝒎𝒆𝒏𝒕 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒊. 𝒊. 𝒅 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝓝(𝟎, 𝟏)
𝟎 , 𝒔𝒊 𝒕 ≠ 𝒔
⇒ 𝑪𝒐𝒗(𝒖𝒕 , 𝒖𝒔 ) = {
𝑽(𝒖𝒕 ) = 𝟏 , 𝒔𝒊 𝒕 = 𝒔
′
𝑽(𝚫𝒚𝒕 ) = 𝑽(𝒖
⏟ 𝒕 ) + 𝑽(𝒖
⏟ 𝒕−𝟏 ) − 𝟐 𝑪𝒐𝒗(𝒖
⏟ 𝒕 , 𝒖𝒕−𝟏 ) . 𝑫 𝒐ù 𝑽(𝚫𝒚𝒕 ) = 𝟐
𝟏 𝟏 𝟎

3)

𝑪𝒐𝒗(𝚫𝒚𝒕 , 𝚫𝒚𝒕−𝟏 ) = 𝑪𝒐𝒗(𝒂 + 𝚫𝒖𝒕 , 𝒂 + 𝚫𝒖𝒕−𝟏 ) = 𝑪𝒐𝒗(𝚫𝒖𝒕 , 𝚫𝒖𝒕−𝟏 ) = 𝑪𝒐𝒗((𝒖𝒕 − 𝒖𝒕−𝟏 ), (𝒖𝒕−𝟏 − 𝒖𝒕−𝟐 ))

𝑪𝒐𝒗(𝒖𝒕 , 𝒖𝒕−𝟏 ) − ⏟
=⏟ 𝑪𝒐𝒗(𝒖𝒕 , 𝒖𝒕−𝟐 ) − ⏟
𝑪𝒐𝒗(𝒖𝒕−𝟏 , 𝒖𝒕−𝟏 ) + ⏟
𝑪𝒐𝒗(𝒖𝒕−𝟏 , 𝒖𝒕−𝟐 )
𝟎 𝟎 𝑽(𝒖𝒕 )=𝟏 𝟎

𝒑𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 , 𝑪𝒐𝒗(𝚫𝒚𝒕 , 𝚫𝒚𝒕−𝟏 ) = −𝟏

𝑶𝒓 𝑽(𝚫𝒚𝒕−𝟏 ) = 𝑽(𝒂 + (𝒖𝒕−𝟏 − 𝒖𝒕−𝟐 )) = 𝑽(𝒖𝒕−𝟏 − 𝒖𝒕−𝟐 ) = ⏟
𝑽(𝒖𝒕−𝟏 ) + ⏟
𝑽(𝒖𝒕−𝟐 ) − 𝟐 ⏟
𝑪𝒐𝒗(𝒖𝒕−𝟏 , 𝒖𝒕−𝟐 ) = 𝟐
𝟏 𝟏 𝟎

𝑪𝒐𝒗(𝚫𝒚𝒕 , 𝚫𝒚𝒕−𝟏 ) −𝟏
𝝆𝜟𝒚𝒕 ,𝜟𝒚𝒕−𝟏 = = . 𝑫′ 𝒐ù 𝝆𝚫𝒚𝒕,𝚫𝒚𝒕−𝟏 = − 𝟏⁄𝟐
√𝑽(𝚫𝒚𝒕 )𝑽(𝚫𝒚𝒕−𝟏 ) √𝟐 × 𝟐

3)

𝑪𝒐𝒗(𝚫𝒚𝒕 , 𝚫𝒚𝒕−𝟐 ) = 𝑪𝒐𝒗(𝒂 + 𝚫𝒖𝒕 , 𝒂 + 𝚫𝒖𝒕−𝟐 ) = 𝑪𝒐𝒗(𝚫𝒖𝒕 , 𝚫𝒖𝒕−𝟐 ) = 𝑪𝒐𝒗((𝒖𝒕 − 𝒖𝒕−𝟏 ), (𝒖𝒕−𝟐 − 𝒖𝒕−𝟑 ))

𝑪𝒐𝒗(𝒖𝒕 , 𝒖𝒕−𝟐 ) − ⏟
=⏟ 𝑪𝒐𝒗(𝒖𝒕 , 𝒖𝒕−𝟑 ) − ⏟
𝑪𝒐𝒗(𝒖𝒕−𝟏 , 𝒖𝒕−𝟐 ) + ⏟
𝑪𝒐𝒗(𝒖𝒕−𝟏 , 𝒖𝒕−𝟑 )
𝟎 𝟎 𝟎 𝟎

𝒑𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 , 𝑪𝒐𝒗(𝚫𝒚𝒕 , 𝚫𝒚𝒕−𝟐 ) = 𝟎 𝒆𝒕 𝝆𝚫𝒚𝒕,𝚫𝒚𝒕−𝟐 = 𝟎

5)
i. 𝒚𝒕 = 𝒍𝒏 𝒙𝒕 = 𝒂𝒕 + 𝒖𝒕 ⟹ 𝒚 ̂
̂𝒕 = 𝒍𝒏 ̂𝒕
𝒙𝒕 = 𝒂

̂ 𝒕 = 𝒚𝒕 − 𝒚
𝒖 ̂𝒕
𝑳𝒂 𝒎é𝒕𝒉𝒐𝒅𝒆 𝒅𝒆𝒔 𝒎𝒐𝒊𝒏𝒅𝒓𝒆𝒔 𝒄𝒂𝒓𝒓é𝒔 𝒐𝒓𝒅𝒊𝒏𝒂𝒊𝒓𝒆𝒔 𝒄𝒐𝒏𝒔𝒊𝒔𝒕𝒆 à 𝒎𝒊𝒏𝒊𝒎𝒊𝒔𝒆𝒓 𝒍𝒂 𝒔𝒐𝒎𝒎𝒆 𝒅𝒆𝒔 𝒄𝒂𝒓𝒓é𝒔
𝑻 𝑻 𝑻

̂ 𝟐𝒕 = 𝚿(𝒂
𝒅𝒆𝒔 𝒓é𝒔𝒊𝒅𝒖𝒔 : ∑ 𝒖 ̂) ⟹ 𝚿(𝒂 ̂𝒕 )𝟐 = ∑(𝒚𝒕 − 𝒂
̂) = ∑(𝒚𝒕 − 𝒚 ̂𝒕)𝟐
𝒕=𝟏 𝒕=𝟏 𝒕=𝟏
𝑻 𝑻 𝑻
′ (𝒂) ′ (𝒚
(𝒚𝒕 − 𝒂
𝚿 ̂ = ∑𝟐⏟ ̂𝒕) 𝒕 ̂𝒕) = ∑ −𝟐𝒕 (𝒚𝒕 − 𝒂
−𝒂 ̂ 𝒕𝟐 )
̂𝒕) = ∑(−𝟐𝒕𝒚𝒕 + 𝟐𝒂
𝒕=𝟏 −𝒕 𝒕=𝟏 𝒕=𝟏
𝑻
′′ (𝒂)
𝚿 ̂ = ∑ 𝟐𝒕𝟐
𝒕=𝟏

323 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝑻

̂𝒕𝟐 ) = 𝟎 (𝟏)
∑(−𝟐𝒕𝒚𝒕 + 𝟐𝒂
′ (𝒂)
𝚿 ̂ =𝟎 𝒕=𝟏
̂ 𝒔𝒐𝒍𝒖𝒕𝒊𝒐𝒏 𝒅𝒆 (𝑺) { ′′
𝒂 ⇔ 𝑻
𝚿 (𝒂̂) > 𝟎
∑ 𝟐𝒕𝟐 > 𝟎 (𝟐)
{ 𝒕=𝟏
𝑻 𝑻 𝑻 𝑻 𝑻
𝟐)
∑𝑻𝒕=𝟏 𝒕𝒚𝒕
𝟐 𝟐
(𝟏): ∑(𝟐𝒕𝒚𝒕 − 𝟐𝒂
̂𝒕 ̂∑𝒕 = 𝟎 ⇔ 𝒂
= 𝟎 ⇔ 𝟐 ∑ 𝒕𝒚𝒕 − 𝟐𝒂 ̂ ∑ 𝒕 = ∑ 𝒕𝒚𝒕 ⇔ 𝒂
̂= 𝑻 𝟐
∑𝒕=𝟏 𝒕
𝒕=𝟏 𝒕=𝟏 𝒕=𝟏 𝒕=𝟏 𝒕=𝟏

𝑻 𝑻 𝑻
𝟐
̂ = (∑ 𝒕𝒚𝒕 )⁄(∑ 𝒕𝟐 )
(𝟐): ∑ 𝟐𝒕 > 𝟎 é𝒗𝒊𝒅𝒆𝒏𝒕𝒆 . 𝑫’𝒐ù 𝒂
𝒕=𝟏 𝒕=𝟏 𝒕=𝟏

ii. ̂ = (∑𝑻𝒕=𝟏 𝒕 𝒍𝒏 𝒙𝒕 )⁄(∑𝑻𝒕=𝟏 𝒕𝟐 )

𝒂
6)
i. 𝒚𝒕 = 𝒍𝒏 𝒙𝒕 = 𝒖𝒕 ⇔ 𝒙𝒕 = 𝒆𝒖𝒕

𝟏 𝒖𝟐
−
𝒂𝒗𝒆𝒄 𝒖𝒕 ↝ 𝓝(𝟎, 𝟏), 𝒅𝒐𝒏𝒄 𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝑼 𝒔𝒆𝒓𝒂 ∶ 𝒇𝑼 (𝒖) = 𝒆 𝟐 𝒐ù 𝒖 ∈ 𝛀𝑼 = ]−∞, +∞[
√𝟐𝝅
𝑵𝒐𝒕𝒐𝒏𝒔 𝑭𝑼 (𝒖) = 𝑷(𝑼 ≤ 𝒖) 𝒆𝒕 𝑭𝑿 (𝒙) = 𝑷(𝑿 ≤ 𝒙), 𝒍𝒆𝒔 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏𝒔 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏𝒔
𝒓𝒆𝒔𝒑𝒆𝒄𝒕𝒊𝒗𝒆𝒔 𝒅𝒆𝒔 𝒗. 𝒂 𝑼 𝒆𝒕 𝑿
𝑭𝑿 (𝒙) = 𝑷(𝑿 ≤ 𝒙) = 𝑷(𝒆𝑼 ≤ 𝒙) 𝒐𝒓 𝑳𝒏 𝒆𝒔𝒕 ↗ 𝒔𝒖𝒓 ]𝟎, +∞[

𝒅𝒐𝒏𝒄 𝑭𝑿 (𝒙) = 𝑷(𝒍𝒏(𝒆𝑼 ) ≤ 𝒍𝒏(𝒙)) ⇔ 𝑭𝑿 (𝒙) = 𝑷(𝑼 ≤ 𝒍𝒏(𝒙)) = 𝑭𝑼 (𝒍𝒏(𝒙))

𝒍𝒏(𝒙)
𝟏 𝒖𝟐
′ −
𝑫 𝒐ù 𝑭𝑿 (𝒙) = 𝑭𝑼 (𝒍𝒏(𝒙)) = ∫ 𝒆 𝟐 𝒅𝒖
√𝟐𝝅
−∞

′ 𝟏
ii. 𝒇𝑿 (𝒙) = 𝑭′𝑿 (𝒙) = (𝒍𝒏(𝒙)) 𝑭′𝑼 (𝒍𝒏(𝒙)) = 𝒙 𝒇𝑼 (𝒍𝒏(𝒙))

𝟐
𝟏 𝟏 −(𝒍𝒏(𝒙))
𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 𝒇𝑿 (𝒙) = 𝒆 𝟐 𝒐ù 𝒙 ∈ 𝛀𝑿 = ]𝟎, +∞[
𝒙 √𝟐𝝅

iii.
𝟏
𝑻 𝑻 𝑻
𝑻
̅ 𝑿 = √∏ 𝒙𝒕 = (∏ 𝒙𝒕 )
𝑮
𝒕=𝟏 𝒕=𝟏

𝟏
𝑻 𝑻 𝑻 𝑻 𝑻
𝟏 𝟏 𝟏
̅ 𝑿 ) = 𝒍𝒏 ((∏ 𝒙𝒕 ) ) = 𝒍𝒏 (∏ 𝒙𝒕 ) = ∑ 𝒍𝒏 𝒙𝒕 = ∑ 𝒖𝒕 . 𝑫′𝒐ù 𝒍𝒏(𝑮
𝒂𝒊𝒏𝒔𝒊, 𝒍𝒏(𝑮 ̅ 𝑿) = 𝑼
̅
𝑻 𝑻 𝑻
𝒕=𝟏 𝒕=𝟏 𝒕=𝟏 𝒕=𝟏

324 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
iv. 𝒍𝒏(𝑮̅ 𝑿) = 𝑼
̅ ⇔𝑮̅ 𝑿 = 𝒆𝑼̅

̅ ̅ 𝜶
̅ 𝑿 )𝜶 ) ⇒ 𝑴𝑼̅ (𝟏) = 𝑬((𝑮
̅ 𝑿 )𝟏 ) = 𝑬(𝑮
̅ 𝑿)
𝒐𝒓 𝑴𝑼̅ (𝜶) = 𝑬(𝒆𝜶𝑼 ) = 𝑬 ((𝒆𝑼 ) ) = 𝑬((𝑮
𝑻
𝟏 𝟏
𝑪𝒂𝒍𝒄𝒖𝒍𝒐𝒏𝒔 𝒅’𝒂𝒃𝒐𝒓𝒅𝒔 𝑴𝑼̅ (𝜶): 𝑴𝑼̅ (𝜶) = 𝑴𝟏 ∑𝑻 (𝜶) = 𝑴∑𝑻 𝒖𝒕 ( 𝜶) = ∏ 𝑴𝒖𝒕 ( 𝜶)
𝒖
𝑻 𝒕=𝟏 𝒕
𝒕=𝟏 𝑻 𝑻
𝒕=𝟏

𝟏 𝟐
𝜶𝟐 𝟏 ( 𝜶)
𝑻 𝜶𝟐
𝒐𝒓 𝑴𝒖𝒕 (𝜶) = 𝒆𝟐 𝒑𝒖𝒊𝒔𝒒𝒖𝒆 𝒖𝒕 ↝ 𝓝(𝟎, 𝟏) ⇒ 𝑴𝒖𝒕 ( 𝜶) = 𝒆 𝟐 = 𝒆 𝟐
𝟐𝑻
𝑻
𝑻 𝑻
𝜶𝟐 𝜶𝟐 𝜶𝟐 𝟏 𝟏
𝑴𝑼̅ (𝜶) = ∏ 𝒆𝟐𝑻𝟐 = (𝒆𝟐𝑻𝟐 ) ̅ 𝑿 ) = 𝒆𝟐𝑻
= 𝒆𝟐𝑻 ⇒ 𝑴𝑼̅ (𝟏) = 𝒆𝟐𝑻 . 𝑫′ 𝒐ù 𝑬(𝑮
𝒕=𝟏

Exercice 6 -Deuxième Partie : (3 points : 1,5+1,5)

ÉNONCÉ
Deuxième Partie :
𝑶𝒏 𝒔𝒖𝒑𝒑𝒐𝒔𝒆 𝒅𝒂𝒏𝒔 𝒄𝒆𝒕𝒕𝒆 𝒑𝒂𝒓𝒕𝒊𝒆, 𝒒𝒖𝒆 𝒍𝒆𝒔 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆𝒔 𝒂 𝒆𝒕 𝒃 𝒔𝒐𝒏𝒕 𝒏𝒖𝒍𝒔 𝒆𝒏 𝒈𝒂𝒓𝒅𝒂𝒏𝒕 𝒍𝒆𝒔 𝒎𝒆𝒎𝒆𝒔
𝒉𝒚𝒑𝒐𝒕𝒉è𝒔𝒆𝒔 𝒔𝒖𝒓 𝒍𝒆𝒔 𝒕𝒆𝒓𝒎𝒆𝒔 𝒅′ 𝒆𝒓𝒓𝒆𝒖𝒓 𝝐𝒕
𝒂𝒗𝒆𝒄 ∶ ▪ 𝒚𝒕 = 𝒂𝒙𝒕 + 𝒃 + 𝝐𝒕 , 𝒕 = 𝟏, 𝟐, … , 𝟐𝟎𝟎
▪ 𝝐𝒕 𝒔𝒐𝒏𝒕 𝒅𝒆𝒔 𝒕𝒆𝒓𝒎𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒊𝒅𝒆𝒏𝒕𝒊𝒒𝒖𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖é𝒔 𝒔𝒆𝒍𝒐𝒏 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆
𝒅′ 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 𝒏𝒖𝒍𝒍𝒆 𝒆𝒕 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝝈𝟐
1) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍’𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 𝒆𝒕 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒅𝒆 𝒚𝟐𝒕
2) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒅𝒆 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒛 = 𝒚𝟐𝒕

Corrigé
1) 𝑶𝒏 𝒂 = 𝒃 = 𝟎 , 𝒑𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆, 𝒚𝒕 = 𝝐𝒕 , 𝒐𝒓 𝝐𝒕 ↝ 𝓝(𝟎, 𝝈𝟐 ) , 𝒅𝒐𝒏𝒄 𝒅𝒆 𝒎ê𝒎𝒆 𝒑𝒐𝒖𝒓 𝒍𝒂 𝒗. 𝒂. 𝒚𝒕

𝑬𝒏 𝒆𝒇𝒇𝒆𝒕 , 𝒚𝒕 ↝ 𝓝(𝟎, 𝝈𝟐 ) ⇒ 𝒚𝒕 ⁄𝝈 ↝ 𝓝(𝟎, 𝟏) ⇒ 𝒚𝟐𝒕 ⁄𝝈𝟐 ↝ 𝝌𝟐 (𝟏)

𝑬(𝒚𝟐𝒕 ⁄𝝈𝟐 ) = 𝟏 𝑬(𝒚𝟐𝒕 )⁄𝝈𝟐 = 𝟏 ′

𝑬(𝒚𝟐𝒕 ) = 𝝈𝟐
𝒂𝒊𝒏𝒔𝒊 ∶ { ⟺{ . 𝑫 𝒐ù {
𝑽(𝒚𝟐𝒕 ⁄𝝈𝟐 ) = 𝟐 𝑽(𝒚𝟐𝒕 )⁄𝝈𝟒 = 𝟐 𝑽(𝒚𝟐𝒕 ) = 𝟐𝝈𝟒

325 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
2)

𝟏 𝒚𝟐
𝟐 ), − 𝒕𝟐
𝒚𝒕 ↝ 𝓝(𝟎, 𝝈 𝒅𝒐𝒏𝒄 𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 𝒔𝒆𝒓𝒂 ∶ 𝒇𝒀 (𝒚𝒕 ) = 𝒆 𝟐𝝈 𝒐ù 𝒚𝒕 ∈ 𝛀𝒀 = ]−∞, +∞[
𝝈√𝟐𝝅
𝑵𝒐𝒕𝒐𝒏𝒔 𝑭𝒀 (𝒚𝒕 ) = 𝑷(𝒀 ≤ 𝒚𝒕 ) 𝒆𝒕 𝑭𝒁 (𝒛𝒕 ) = 𝑷(𝒁 ≤ 𝒛𝒕 ), 𝒍𝒆𝒔 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏𝒔 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏𝒔 𝒓𝒆𝒔𝒑𝒆𝒄𝒕𝒊𝒗𝒆𝒔
𝒅𝒆𝒔 𝒗. 𝒂 𝒀 𝒆𝒕 𝒁,
𝑶𝒏 𝒂 ∶ 𝒁 = 𝒀𝟐 , 𝒑𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 𝛀𝒀 = ]−∞, +∞[ ⇒ 𝛀𝒁 ⊆ [𝟎, +∞[

𝑭𝒁 (𝒛𝒕 ) = 𝑷(𝒁 ≤ 𝒛𝒕 ) = 𝑷(𝒀𝟐 ≤ 𝒛𝒕 ) = 𝑷(−√𝒛𝒕 ≤ 𝒀 ≤ √𝒛𝒕 ) = 𝑭𝒀 (√𝒛𝒕 ) − 𝑭𝒀 (−√𝒛𝒕 )

𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒁 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏 ∶

𝒅𝑭𝒁 (𝒛𝒕 ) 𝒅𝑭𝒀 (√𝒛𝒕 ) 𝒅𝑭𝒀 (−√𝒛𝒕 )

∀𝒛𝒕 ∈ ]𝟎, +∞[ ; 𝒇𝒁 (𝒛𝒕 ) = = −
𝒅𝒛𝒕 𝒅𝒛𝒕 𝒅𝒛𝒕
′ ′ 𝟏 𝟏
∀𝒛𝒕 ∈ ]𝟎, +∞[; 𝒇𝒁 (𝒛𝒕 ) = (√𝒛𝒕 ) 𝒇𝒀 (√𝒛𝒕 ) − (−√𝒛𝒕 ) 𝒇𝒀 (−√𝒛𝒕 ) = 𝒇𝒀 (√𝒛𝒕 ) + 𝒇𝒀 (−√𝒛𝒕 )
𝟐√𝒛𝒕 𝟐√𝒛𝒕

𝑶𝒏 𝒓𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒓𝒂 𝒇𝒂𝒄𝒊𝒍𝒆𝒎𝒆𝒏𝒕 𝒒𝒖𝒆 𝒇𝒀 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒑𝒂𝒊𝒓𝒆 ⇒ 𝒇𝒀 (−√𝒛𝒕 ) = 𝒇𝒀 (√𝒛𝒕 )

𝟐
(√𝒛𝒕 ) 𝒛
− − 𝒕𝟐
𝟏 𝟏 𝒆 𝟐𝝈𝟐 𝒆 𝟐𝝈
𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 , ∀𝒛𝒕 ∈ ]𝟎, +∞[; 𝒇𝒁 (𝒛𝒕 ) = 𝒇𝒀 (√𝒛𝒕 ) = =
√𝒛𝒕 √𝒛𝒕 𝝈√𝟐𝝅 √𝝅√𝟐𝝈𝟐 √𝒛𝒕
𝒛 𝟏 𝒛
− 𝒕𝟐 − 𝒕𝟐
𝒆 𝟐𝝈 𝒛𝒕 𝟐−𝟏 𝒆 𝟐𝝈
∀𝒛𝒕 ∈ ]𝟎, +∞[; 𝒇𝒁 (𝒛𝒕 ) = 𝟏 𝟏 = 𝟏
√𝝅(𝟐𝝈𝟐 )𝟐 𝒛𝒕 𝟐 √𝝅(𝟐𝝈𝟐 )𝟐
𝟏 𝒛
− 𝒕𝟐
[𝒛𝒕 𝟐−𝟏 ] [𝒆 𝟐𝝈 ]

𝑫’𝒐ù 𝒍𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒅𝒆 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒁 = 𝒀𝟐 ∶ 𝒇𝒁 (𝒛𝒕 ) = 𝟏 , 𝒔𝒊 𝒛𝒕 > 𝟎

√𝝅(𝟐𝝈𝟐 )𝟐
{𝟎, 𝒔𝒊 𝒏𝒐𝒏
𝑶𝒏 𝒑𝒐𝒖𝒓𝒓𝒂 𝒓𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒓 𝒒𝒖𝒆 𝒍𝒂 𝒗. 𝒂. 𝒁 𝒆𝒔𝒕 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖é𝒆 𝒔𝒖𝒊𝒗𝒂𝒏𝒕 𝒍𝒂 𝒍𝒐𝒊 𝑮𝒂𝒎𝒎𝒂 𝒅𝒆 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆
𝟏 𝟏
𝒆𝒕 𝟐𝝈𝟐 ∶ 𝜞 ( , 𝟐𝝈𝟐 )
𝟐 𝟐
𝒛
𝟏 − 𝒕𝟐
𝟏 𝒛 𝒛𝒕
𝟏
( −𝟏)
𝟐 𝒛 𝒛𝒕
𝟏
( −𝟏)
𝟐 𝒛 [𝒛𝒕 (𝟐−𝟏) ] [𝒆 𝟐𝝈 ]
− 𝒕𝟐 − 𝒕𝟐 − 𝒕𝟐
𝒇𝒁 (𝒛𝒕 ) = 𝒆 𝟐𝝈 = 𝒆 𝟐𝝈 = 𝟏 𝒆 𝟐𝝈 = 𝟏
𝝈√𝟐𝝅𝒛𝒕 √𝝅√𝟐𝝈𝟐 √𝝅(𝟐𝝈𝟐 )𝟐 𝜞(𝟏/𝟐)(𝟐𝝈𝟐 )𝟐
𝟏
𝒄𝒂𝒓 𝒐𝒏 𝒂 ∶ 𝜞 ( ) = √𝝅
𝟐
𝑶𝒏 𝒑𝒐𝒖𝒓𝒓𝒂 𝒂𝒖𝒔𝒔𝒊 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒍𝒂 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝑮𝒂𝒎𝒎𝒂 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒁 𝒅′ 𝒖𝒏𝒆 𝒂𝒖𝒕𝒓𝒆 𝒎𝒂𝒏𝒊è𝒓𝒆, 𝒄‑à‑𝒅.
𝒆𝒏 𝒊𝒏𝒕𝒓𝒐𝒅𝒖𝒊𝒔𝒂𝒏𝒕 𝒍𝒂 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒅𝒆 𝒌𝒉𝒊‑𝒅𝒆𝒖𝒙 ∶
𝒐𝒏 𝒂 ∶ 𝒚𝟐𝒕 ⁄𝝈𝟐 ↝ 𝝌𝟐 (𝟏) , 𝒅𝒐𝒏𝒄 𝒛𝒕 ⁄𝝈𝟐 ↝ 𝝌𝟐 (𝟏) , 𝒐𝒓 𝜞(𝟏⁄𝟐 , 𝟐) ≡ 𝝌𝟐 (𝟏) , 𝒅𝒐𝒏𝒄 𝒛𝒕 ⁄𝝈𝟐 ↝ 𝜞(𝟏⁄𝟐 , 𝟐)

326 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒐𝒓 𝒛𝒕 ⁄𝝈𝟐 ↝ 𝜞(𝟏⁄𝟐 , 𝟐) ⇒ 𝝈𝟐 ( 𝒛𝒕 ⁄𝝈𝟐 ) ↝ 𝜞(𝟏⁄𝟐 , 𝟐𝝈𝟐 ), 𝒑𝒖𝒊𝒔𝒒𝒖𝒆 𝝈𝟐 > 𝟎
𝟏 𝒛
− 𝒕𝟐
(𝒛𝒕 𝟐−𝟏 ) (𝒆 𝟐𝝈 )

𝒅′ 𝒐ù 𝒐𝒏 𝒓𝒆𝒕𝒓𝒐𝒖𝒗𝒆 ∶ 𝒛𝒕 ↝ 𝜞(𝟏⁄𝟐 , 𝟐𝝈𝟐 ) 𝒆𝒕 𝒇𝒁 (𝒛𝒕 ) = 𝟏 , 𝒔𝒊 𝒛𝒕 > 𝟎

√ 𝝅(𝟐𝝈𝟐 )𝟐
{ 𝟎, 𝒔𝒊 𝒏𝒐𝒏

Exercice 7 – Partie 1 : (5 points : 1+1+1+2)

ÉNONCÉ
Les deux parties de cet exercice sont indépendantes.
𝑪𝒐𝒏𝒔𝒊𝒅é𝒓𝒐𝒏𝒔 𝑿 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒂𝒚𝒂𝒏𝒕 𝒖𝒏𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒆𝒙𝒑𝒐𝒏𝒆𝒏𝒕𝒊𝒆𝒍𝒍𝒆 𝒅𝒆 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆 𝜽 ,
𝜽𝒆−𝜽𝒙 , 𝒑𝒐𝒖𝒓 𝒙 ≥ 𝟎
𝒅é𝒇𝒊𝒏𝒊𝒆 𝒑𝒂𝒓 ∶ 𝒇(𝒙) = { ; 𝜽 𝒆𝒔𝒕 𝒖𝒏 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆 𝒑𝒐𝒔𝒊𝒕𝒊𝒇.
𝟎 , 𝒔𝒊 𝒏𝒐𝒏

Première Partie :
1) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝑿
2) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒆𝒏 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝜽 𝒍𝒂 𝒎é𝒅𝒊𝒂𝒏𝒆 𝒅𝒆 𝑿
3) 𝑶𝒏 𝒅é𝒇𝒊𝒏𝒊𝒕 𝒁 = ⌊𝑿⌋ 𝒒𝒖𝒊 𝒅é𝒔𝒊𝒈𝒏𝒆 𝒍𝒂 𝒑𝒂𝒓𝒕𝒊𝒆 𝒆𝒏𝒕𝒊è𝒓𝒆 𝒅𝒆 𝑿 ∶ 𝒄𝒆 𝒒𝒖𝒊 𝒔𝒊𝒈𝒏𝒊𝒇𝒊𝒆 𝒒𝒖𝒆 𝒁 𝒆𝒔𝒕 𝒍𝒆
𝒑𝒍𝒖𝒔 𝒈𝒓𝒂𝒏𝒅 𝒏𝒐𝒎𝒃𝒓𝒆 𝒆𝒏𝒕𝒊𝒆𝒓 𝒊𝒏𝒇é𝒓𝒊𝒆𝒖𝒓 à 𝑿
i. 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒅𝒆 𝒁
ii. 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍’𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 𝒅𝒆 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒁

Corrigé
1)

∙ 𝑺𝒊 𝒙 < 𝟎 , 𝒂𝒍𝒐𝒓𝒔 𝑭𝑿 (𝒙) = 𝑷(𝑿 ≤ 𝒙) = 𝟎

𝒙 𝒙 𝟎
𝟎
∙ 𝑺𝒊 𝒙 ≥ 𝟎 , 𝒂𝒍𝒐𝒓𝒔 𝑭𝑿 (𝒙) = 𝑷(𝑿 ≤ 𝒙) = ∫ 𝒇𝑿 (𝒕)𝒅𝒕 = ∫ 𝜽𝒆−𝜽𝒕 𝒅𝒕 = ∫(−𝜽𝒕)′ 𝒆−𝜽𝒕 𝒅𝒕 = [𝒆−𝜽𝒕 ]𝒙
𝟎 𝟎 𝒙

′ 𝟎, 𝒔𝒊 𝒙 < 𝟎
𝑺𝒊 𝒙 ≥ 𝟎 , 𝒂𝒍𝒐𝒓𝒔 𝑭𝑿 (𝒙) = 𝟏 − 𝒆−𝜽𝒙 . 𝑫 𝒐ù 𝑭𝑿 (𝒙) = {
𝟏 − 𝒆−𝜽𝒙 , 𝒔𝒊 𝒙 ≥ 𝟎

2) 𝑷(𝑿 ≤ 𝑴𝒆 ) = 𝑭𝑿 (𝑴𝒆 ) = 𝟏⁄𝟐 ⟺ 𝟏 − 𝒆−𝜽𝑴𝒆 = 𝟏⁄𝟐 ⟺ 𝒆−𝜽𝑴𝒆 = 𝟏⁄𝟐 ⟺ −𝜽𝑴𝒆 = 𝐥𝐧(𝟏⁄𝟐)

327 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝐥𝐧(𝟐)
⟺ −𝜽𝑴𝒆 = − 𝐥𝐧(𝟐) . 𝑫’𝒐ù 𝑴𝒆 =
𝜽
3)
i. 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒅𝒆 𝒁 ∶

𝑻𝒐𝒖𝒕 𝒅’𝒂𝒃𝒐𝒓𝒅 𝒁(𝛀) = ℕ, 𝒄𝒆 𝒒𝒖𝒊 𝒔𝒊𝒈𝒏𝒊𝒇𝒊𝒆 𝒒𝒖𝒆 𝒁 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒅𝒊𝒔𝒄𝒓è𝒕𝒆.

𝑷(𝒁 = 𝒛) = 𝑷(⌊𝑿⌋ = 𝒛) = 𝑷(𝒛 ≤ 𝑿 < 𝒛 + 𝟏) = 𝑭𝑿 (𝒛 + 𝟏) − 𝑭𝑿 (𝒛) = [𝟏 − 𝒆−𝜽(𝒛+𝟏) ] − [𝟏 − 𝒆−𝜽𝒛 ]

𝑷(𝒁 = 𝒛) = 𝒆−𝜽𝒛 − 𝒆−𝜽𝒛−𝜽 = 𝒆−𝜽𝒛 (𝟏 − 𝒆−𝜽 )

𝑵𝒐𝒕𝒐𝒏𝒔 𝒑 = 𝟏 − 𝒆−𝜽 , 𝒑𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 𝟏 − 𝒑 = 𝒆−𝜽 .

𝑶𝒏 𝒗é𝒓𝒊𝒇𝒊𝒆 𝒂𝒖𝒔𝒔𝒊 𝒒𝒖𝒆 𝜽 ∈ ]𝟎, 𝟏[ ∶
𝜽 > 𝟎 ⟺ −𝜽 < 𝟎 ⟺ 𝟎 < 𝒆−𝜽 < 𝟏 ⟺ −𝟏 < −𝒆−𝜽 < 𝟎 ⟺ 𝟎 < ⏟
𝟏 − 𝒆−𝜽 < 𝟏
𝒑
𝒛
𝑷(𝒁 = 𝒛) = (𝒆−𝜽 ) (𝟏 − 𝒆−𝜽 ) = (𝟏 − 𝒑)𝒛 𝒑 ; 𝒂𝒗𝒆𝒄 𝒑 = 𝟏 − 𝒆−𝜽
𝒛
(𝒆−𝜽 ) (𝟏 − 𝒆−𝜽 ) , 𝒔𝒊 𝒛 ∈ ℕ
𝑷(𝒁 = 𝒛) = {
𝟎 , 𝒔𝒊 𝒏𝒐𝒏

iii. 𝑫é𝒎𝒐𝒏𝒕𝒓𝒐𝒏𝒔 𝒒𝒖𝒆 𝒍𝒂 𝒗. 𝒂. 𝑻 = 𝒁 + 𝟏 ↝ 𝓖 (𝟏 − 𝒆−𝜽 ) ∶

𝑻=𝒁+𝟏
∙{ ⇒ 𝑻(𝛀) = ℕ∗
𝒁(𝛀) = ℕ

∙ 𝑭𝑻 (𝒕) = 𝑷(𝑻 ≤ 𝒕) = 𝑷(𝒁 + 𝟏 ≤ 𝒕) = 𝑷(𝒁 ≤ 𝒕 − 𝟏) ⟺ 𝑭𝑻 (𝒕) = 𝑭𝒁 (𝒕 − 𝟏)

𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒁 :

+∞

∙ ∀𝒛 ≥ 𝟎, 𝑭𝒁 (𝒛) = 𝑷(𝒁 ≤ 𝒛) = 𝟏 − 𝑷(𝒁 > 𝒛) = 𝟏 − 𝑷(𝒁 ≥ 𝒛 + 𝟏) = 𝟏 − ∑ 𝑷(𝒁 = 𝒖)

𝒖=𝒛+𝟏
+∞ +∞
−𝜽 𝒛 𝒛
∀𝒛 ≥ 𝟎, 𝑭𝒁 (𝒛) = 𝟏 − ∑ (𝒆 ) (𝟏 − 𝒆−𝜽 ) = 𝟏 − (𝟏 − 𝒆−𝜽 ) ∑ (𝒆−𝜽 )
𝒖=𝒛+𝟏 𝒖=𝒛+𝟏
+∞ 𝒛+𝟏
−𝜽 −𝜽 𝒛
(𝒆−𝜽 )
𝒐𝒏 𝒂 𝜽 > 𝟎, 𝒄𝒆 𝒒𝒖𝒊 𝒅𝒐𝒏𝒏𝒆 ∶ 𝒆 ∈ ]𝟎, 𝟏[ 𝒂𝒊𝒏𝒔𝒊 ∑ (𝒆 ) =
𝟏 − 𝒆−𝜽
𝒖=𝒛+𝟏
𝒛+𝟏
−𝜽
(𝒆−𝜽 ) 𝒛+𝟏
∀𝒛 ≥ 𝟎, 𝑭𝒁 (𝒛) = 𝟏 − (𝟏 − 𝒆 )[ −𝜽
] = 𝟏 − (𝒆−𝜽 )
𝟏−𝒆

𝟎 , 𝒔𝒊 𝒛 < 𝟎
𝑭𝒁 (𝒛) = { 𝒛+𝟏
𝟏 − (𝒆−𝜽 ) , 𝒔𝒊 𝒛 ≥ 𝟎
(𝒕−𝟏)+𝟏
∙ ∀𝒕 ≥ 𝟏, 𝑭𝑻 (𝒕) = 𝑭𝒁 (𝒕 − 𝟏) = 𝟏 − (𝒆−𝜽 )

328 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝟎 , 𝒔𝒊 𝒕 < 𝟏
𝑭𝑻 (𝒕) = { 𝒕
𝟏 − (𝒆−𝜽 ) , 𝒔𝒊 𝒕 ≥ 𝟏

∙ 𝑶𝒏 𝒂 ∶ ∀𝒕 ∈ ℕ∗ , 𝑷(𝑻 = 𝒕) = 𝑷(𝑻 ≤ 𝒕) − 𝑷(𝑻 ≤ 𝒕 − 𝟏) = 𝑭𝑻 (𝒕) − 𝑭𝑻 (𝒕 − 𝟏)

𝒕 𝒕−𝟏 𝒕−𝟏 𝒕 𝒕−𝟏
∀𝒕 ∈ ℕ∗ , 𝑷(𝑻 = 𝒕) = [𝟏 − (𝒆−𝜽 ) ] − [𝟏 − (𝒆−𝜽 ) ] = (𝒆−𝜽 ) − (𝒆−𝜽 ) = (𝒆−𝜽 ) (𝟏 − 𝒆−𝜽 )

−𝜽 𝒕−𝟏
𝑷(𝑻 = 𝒕) = {(𝒆 ) (𝟏 − 𝒆−𝜽 ) = (𝟏 − 𝒑)𝒕−𝟏 𝒑 , 𝒔𝒊 𝒕 ∈ ℕ∗ 𝒂𝒗𝒆𝒄 𝒑 = (𝟏 − 𝒆−𝜽 ) ∈ ]𝟎, 𝟏[
𝟎 , 𝒔𝒊 𝒏𝒐𝒏

𝑫′ 𝒐ù 𝑻 ↝ 𝓖 (𝟏 − 𝒆−𝜽 )

𝟏
𝑬𝒏 𝒆𝒇𝒇𝒆𝒕 𝑬(𝑻) =
𝟏 − 𝒆−𝜽
𝟏 𝟏
𝑶𝒓 𝑻 = 𝒁 + 𝟏 , 𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 𝑬(𝒁 + 𝟏) = ⟺ 𝑬(𝒁) + 𝟏 =
𝟏 − 𝒆−𝜽 𝟏 − 𝒆−𝜽
𝟏 𝟏 − 𝟏 + 𝒆−𝜽 𝒆−𝜽 𝟏 𝟏
⟺ 𝑬(𝒁) = −𝜽
− 𝟏 = −𝜽
= −𝜽
= 𝜽 . 𝑫′ 𝒐ù 𝑬(𝒁) = 𝜽
𝟏−𝒆 𝟏−𝒆 𝟏−𝒆 𝒆 −𝟏 𝒆 −𝟏

Exercice 8 : (7 points : 1 point par question)

ÉNONCÉ
𝑳𝒂 𝒑𝒆𝒓𝒕𝒆 𝒄𝒂𝒖𝒔é𝒆 𝒑𝒂𝒓 𝒖𝒏 𝒂𝒄𝒄𝒊𝒅𝒆𝒏𝒕 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝑿 𝒂𝒚𝒂𝒏𝒕 𝒑𝒐𝒖𝒓 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆
𝒄
𝒑𝒐𝒖𝒓 𝒙 ≥ 𝟏
𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒑𝒂𝒓 : 𝒇𝑿 (𝒙) = {𝒙𝟓 ; 𝒐ù 𝒄 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒄𝒐𝒔𝒕𝒂𝒏𝒕𝒆.
𝟎, 𝒔𝒊 𝒏𝒐𝒏
1) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒗𝒂𝒍𝒆𝒖𝒓 𝒅𝒆 𝒍𝒂 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆 𝒄
2) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍’𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 𝒅𝒆 𝑿
3) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝑿. 𝑬𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒍𝒂 𝒎é𝒅𝒊𝒂𝒏𝒆 𝒅𝒆 𝑿.
4) 𝑳𝒂 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒅𝒆 𝑿 𝒆𝒔𝒕‑ 𝒆𝒍𝒍𝒆 𝒔𝒚𝒎é𝒕𝒓𝒊𝒒𝒖𝒆 ? 𝑱𝒖𝒔𝒕𝒊𝒇𝒊𝒆𝒓 𝒗𝒐𝒕𝒓𝒆 𝒓é𝒑𝒐𝒏𝒔𝒆.
5) 𝑰𝒏𝒕𝒆𝒓𝒑𝒓é𝒕𝒆𝒓 𝒍𝒂 𝒗𝒂𝒍𝒆𝒖𝒓 𝒅𝒆 𝑩
6) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒅𝒆 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒀
𝟕
7) 𝑸𝒖𝒆𝒍𝒍𝒆 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 𝒇𝒂𝒖𝒕‑ 𝒊𝒍 𝒊𝒎𝒑𝒐𝒔𝒆𝒓 𝒔𝒖𝒓 𝒍𝒂 𝒗𝒂𝒍𝒆𝒖𝒓 𝒅𝒆 𝑩 𝒑𝒐𝒖𝒓 𝒒𝒖𝒆 𝑬(𝒀) = 𝟔 ?
𝑪𝒐𝒎𝒎𝒆𝒏𝒕𝒆𝒓 𝒄𝒆 𝒓é𝒔𝒖𝒍𝒕𝒂𝒕

329 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
Corrigé
1)
+∞
∫ 𝒇𝑿 (𝒙)𝒅𝒙 = 𝟏 (𝟏)
𝒇𝑿 (𝒙)𝒆𝒔𝒕 𝒖𝒏𝒆 𝒅. 𝒅. 𝒑 ⟺ { 𝟏
𝒇𝑿 (𝒙) ≥ 𝟎 , ∀𝒙 ∈ [𝟏, +∞[ (𝟐)
+∞ +∞
𝒄 𝒄 𝟏 +∞ 𝒄
(𝟏): ∫ 𝒇𝑿 (𝒙)𝒅𝒙 = 𝟏 ⟺ ∫ 𝒅𝒙 = 𝟏 ⟺ − [ ] = 𝟏 ⟺ =𝟏⟺𝒄=𝟒
𝟏 𝟏 𝒙𝟓 𝟒 𝒙𝟒 𝟏 𝟒
𝟒
(𝟐): 𝒇𝑿 (𝒙) = ≥ 𝟎 , ∀𝒙 ∈ [𝟏, +∞[
𝒙𝟓
𝟒
𝑫′ 𝒐ù, 𝒇𝑿 (𝒙) = {𝒙𝟓 𝒑𝒐𝒖𝒓 𝒙 ≥ 𝟏
𝟎, 𝒔𝒊 𝒏𝒐𝒏

2)
+∞ +∞
𝟒 𝟒 𝟏 +∞ 𝟒
𝑬(𝑿) = ∫ 𝒙𝒇𝑿 (𝒙)𝒅𝒙 = ∫ 𝒅𝒙 = − [ ] = − [𝟎 − 𝟏]
𝟏 𝟏 𝒙𝟒 𝟑 𝒙𝟑 𝟏 𝟑

𝟒
𝑬(𝑿) = ≅ 𝟏, 𝟑𝟑
𝟑
3)
𝟎, 𝒔𝒊 𝒙 < 𝟏
𝒙
▪𝑭𝑿 (𝒙) = 𝑷(𝑿 ≤ 𝒙) {
∫ 𝒇𝑿 (𝒕)𝒅𝒕 , 𝒔𝒊 𝒙 ≥ 𝟏
𝟏

𝟒 𝟒 𝟏 𝒙 𝒙
𝟏
𝑷𝒐𝒖𝒓 𝒕𝒐𝒖𝒕 𝒙 ≥ 𝟏, 𝒐𝒏 𝒂 ∶ 𝑭𝑿 (𝒙) = ∫ 𝟓 𝒅𝒕 = − [ 𝟒 ] = 𝟏 − 𝟒
𝟏 𝒕 𝟒 𝒕 𝟏 𝒙

𝟎, 𝒔𝒊 𝒙 < 𝟏
′ 𝟏
𝑫 𝒐ù, 𝑭𝑿 (𝒙) = 𝑷(𝑿 ≤ 𝒙) {
𝟏 − 𝟒 , 𝒔𝒊 𝒙 ≥ 𝟏
𝒙
𝟏
▪𝑴𝒆 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒔𝒐𝒍𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍’é𝒒𝒖𝒂𝒕𝒊𝒐𝒏 ∶ 𝑭𝑿 (𝒙) = 𝒔𝒖𝒓 𝒍′ 𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆[𝟏, +∞[
𝟐
𝟏 𝟏 𝟏 𝟏 𝟏 𝟒
𝑶𝒓, 𝑭𝑿 (𝒙) = ⟺ 𝟏 − 𝟒 = ⟺ 𝟒 = ⟺ 𝒙𝟒 = 𝟐 ⟺ 𝒙 = √𝟐
𝟐 𝒙 𝟐 𝒙 𝟐
𝟒
𝑫′ 𝒐ù, 𝑴𝒆 = √𝟐 ≅ 𝟏, 𝟏𝟗

4)
𝑴𝒆 ≠ 𝑬(𝑿), 𝒅𝒐𝒏𝒄 𝒍𝒂 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒏’𝒆𝒔𝒕 𝒑𝒂𝒔 𝒔𝒚𝒎é𝒕𝒓𝒊𝒒𝒖𝒆
𝟑
′ 𝟑 𝟑 𝟐 𝟐
𝑬𝒏 𝒄𝒂𝒍𝒄𝒖𝒍𝒂𝒏𝒕 𝒍𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅 𝒂𝒔𝒚𝒎é𝒕𝒓𝒊𝒆 𝜸𝟏 = 𝝁𝟑 ⁄𝝈 = 𝑬 [(𝑿 − 𝑬(𝑿)) ]⁄[𝑬 ((𝑿 − 𝑬(𝑿)) )] ,

𝒐𝒏 𝒕𝒓𝒐𝒖𝒗𝒆𝒓𝒂 𝜸𝟏 ≠ 𝟎 𝒆𝒕 𝒐𝒏 𝒑𝒆𝒖𝒕 𝒂𝒇𝒇𝒊𝒓𝒎𝒆𝒓 𝒒𝒖𝒆 𝒍𝒂 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒏′ 𝒆𝒔𝒕𝒑𝒂𝒔 𝒔𝒚𝒎é𝒕𝒓𝒊𝒒𝒖𝒆.

5) 𝑩 𝒆𝒔𝒕 𝒍𝒆 𝒑𝒍𝒂𝒇𝒐𝒏𝒅 𝒍𝒂 𝒔𝒐𝒎𝒎𝒆 𝒎𝒂𝒙𝒊𝒎𝒖𝒎 𝒅𝒆 𝒓𝒆𝒎𝒃𝒐𝒖𝒓𝒔𝒆𝒎𝒆𝒏𝒕 𝒑𝒓é𝒗𝒖𝒆 𝒂𝒖 𝒄𝒐𝒏𝒕𝒓𝒂𝒕

330 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒅’𝒂𝒔𝒔𝒖𝒓𝒂𝒏𝒄𝒆. 𝑬𝒏 𝒄𝒂𝒔 𝒅𝒆 𝒔𝒊𝒏𝒊𝒔𝒕𝒓𝒆, 𝒍’𝒂𝒔𝒔𝒖𝒓𝒆𝒖𝒓 𝒔’𝒆𝒏𝒈𝒂𝒈𝒆 à 𝒊𝒏𝒅𝒆𝒎𝒏𝒊𝒔𝒆𝒓 𝒍’𝒂𝒔𝒔𝒖𝒓é à 𝒉𝒂𝒖𝒕𝒆𝒖𝒓
𝒅𝒆 𝒄𝒆𝒕𝒕𝒆 𝒔𝒐𝒎𝒎𝒆 𝒎𝒂𝒊𝒔 𝒑𝒂𝒔 𝒂𝒖‑ 𝒅𝒆𝒍à.
𝑪𝒆 𝒑𝒍𝒂𝒇𝒐𝒏𝒅 𝒅𝒆 𝒈𝒂𝒓𝒂𝒏𝒕𝒊𝒆 𝒆𝒔𝒕 𝒇𝒊𝒙é 𝒅𝒂𝒏𝒔 𝒏𝒐𝒕𝒓𝒆 𝒄𝒂𝒔 𝒑𝒂𝒓 𝒔𝒊𝒏𝒊𝒔𝒕𝒓𝒆 (𝒑𝒆𝒓𝒕𝒆).
6)
𝑿, 𝒑𝒐𝒖𝒓 𝒙 < 𝑩
𝒀={ ⟹ 𝒀(𝛀) = [𝟏, 𝑩[ ∪ {𝑩} = [𝟏, 𝑩]
𝑩, 𝒑𝒐𝒖𝒓 𝒙 ≥ 𝑩
𝑬𝒏 𝒆𝒇𝒇𝒆𝒕, 𝒀 = 𝐦𝐢𝐧(𝑿, 𝑩)
▪𝑭𝒀 (𝒚) = 𝑷(𝒀 ≤ 𝒚) = 𝑷(𝐦𝐢𝐧(𝑿, 𝑩) ≤ 𝒚) = 𝟏 − 𝑷(𝐦𝐢𝐧(𝑿, 𝑩) > 𝒚) = 𝟏 − 𝑷(𝑿 > 𝒚 , 𝑩 > 𝒚)
𝑩 é𝒕𝒂𝒏𝒕 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒄𝒆𝒓𝒕𝒂𝒊𝒏𝒆 𝒅𝒐𝒏𝒄 𝒆𝒍𝒍𝒆 𝒆𝒔𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆 𝒂𝒗𝒆𝒄 𝒕𝒐𝒖𝒕𝒆 𝒂𝒖𝒕𝒓𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆
𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆, 𝒑𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 ∶
𝑭𝒀 (𝒚) = 𝟏 − 𝑷(𝑿 > 𝒚)𝑷(𝑩 > 𝒚) = 𝟏 − [(𝟏 − 𝑷(𝑿 ≤ 𝒚))(𝟏 − 𝑷(𝑩 ≤ 𝒚))]
= 𝟏 − [(𝟏 − 𝑭𝑿 (𝒚)) (𝟏 − 𝟙[𝑩,+∞[ (𝒚))]
𝟎, 𝒔𝒊 𝒚 < 𝟏
𝑨𝒗𝒆𝒄𝑭𝑿 (𝒚) = 𝑷(𝑿 ≤ 𝒚) { 𝟏
𝟏 − 𝟒 , 𝒔𝒊 𝒚 ≥ 𝟏
𝒚
𝟎, 𝒔𝒊 𝒚 ∈ ]−∞, 𝑩[
𝒆𝒕 𝟙[𝑩,+∞[ (𝒚) { , 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒊𝒏𝒅𝒊𝒄𝒂𝒕𝒓𝒊𝒄𝒆 𝒔𝒖𝒓 𝒍′ 𝒊𝒏𝒕𝒆𝒓𝒗𝒂𝒍𝒍𝒆 [𝑩, +∞[ 𝒒𝒖𝒊 𝒆𝒔𝒕
𝟏 , 𝒔𝒊 𝒚 ∈ [𝑩, +∞[
𝒂𝒖𝒔𝒔𝒊 𝒍𝒂 𝒇. 𝒓. 𝒅𝒆 𝒍𝒂 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒅𝒆 𝑫𝒊𝒓𝒂𝒄 𝒂𝒖 𝒑𝒐𝒊𝒏𝒕 𝑩 ∶ 𝜹𝑩 𝒂𝒚𝒂𝒏𝒕 𝒑𝒐𝒖𝒓 𝒍𝒐𝒊 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é ∶
𝟏, 𝒔𝒊 𝒚 = 𝑩
𝜹𝑩 (𝒚) = 𝑷(𝒀 = 𝒚) = {
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
𝑭𝒀 (𝒚) = 𝟏 − [𝟏 − 𝟙[𝑩,+∞[ (𝒚) − 𝑭𝑿 (𝒚) + 𝑭𝑿 (𝒚)𝟙[𝑩,+∞[ (𝒚)]
𝑭𝒀 (𝒚) = 𝟙[𝑩,+∞[ (𝒚) + 𝑭𝑿 (𝒚) − 𝑭𝑿 (𝒚)𝟙[𝑩,+∞[ (𝒚)
◦ 𝑺𝒊 𝒚 < 𝟏 ⟹ 𝑭𝑿 (𝒚) = 𝟎 𝒆𝒕 𝟙[𝑩,+∞[ (𝒚) = 𝟎, 𝒅𝒐𝒏𝒄, 𝑭𝒀 (𝒚) = 𝟎
𝟏 𝟏
◦ 𝑺𝒊 𝟏 ≤ 𝒚 < 𝑩 ⟹ 𝑭𝑿 (𝒚) = 𝟏 − 𝟒 𝒆𝒕 𝟙[𝑩,+∞[ (𝒚) = 𝟎, 𝒅𝒐𝒏𝒄, 𝑭𝒀 (𝒚) = 𝟏 − 𝟒
𝒚 𝒚
𝟏 𝟏 𝟏
◦ 𝑺𝒊 𝒚 ≥ 𝑩 ⟹ 𝑭𝑿 (𝒚) = 𝟏 − 𝟒 𝒆𝒕 𝟙[𝑩,+∞[ (𝒚) = 𝟏, 𝒅𝒐𝒏𝒄, 𝑭𝒀 (𝒚) = 𝟏 + 𝟏 − 𝟒 − (𝟏 − 𝟒 ) = 𝟏
𝑩 𝑩 𝑩
𝟎, 𝒔𝒊 𝒚 < 𝟏
𝟏
𝑶𝒏 𝒐𝒃𝒕𝒊𝒆𝒏𝒕 ∶ 𝑭𝒀 (𝒚) = 𝟏 − 𝟒 , 𝒔𝒊 𝟏 ≤ 𝒚 < 𝑩
𝒚
{ 𝟏, 𝒔𝒊 𝒚 ≥ 𝑩
𝑭𝒀 𝒏’𝒆𝒔𝒕 𝒏𝒊 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆 𝒑𝒂𝒓 𝒎𝒐𝒓𝒄𝒆𝒂𝒖𝒙 (𝒄𝒐𝒎𝒎𝒆 𝒍𝒆𝒔 𝒗. 𝒂. 𝒅𝒊𝒔𝒄𝒓è𝒕𝒆𝒔)𝒏𝒊 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆
(𝒄𝒐𝒎𝒎𝒆 𝒍𝒆𝒔 𝒗. 𝒂. à 𝒅𝒆𝒏𝒔𝒊𝒕é). 𝑼𝒏𝒆 𝒕𝒆𝒍𝒍𝒆 𝒗. 𝒂. 𝒆𝒔𝒕 𝒅𝒊𝒕𝒆 𝒎𝒊𝒙𝒕𝒆.
▪𝒇𝒀 (𝒚) = 𝑭′𝒀 (𝒚) = 𝟙′[𝑩,+∞[ (𝒚) + 𝑭′𝑿 (𝒚) − [𝑭′𝑿 (𝒚)𝟙[𝑩,+∞[ (𝒚) + 𝑭𝑿 (𝒚)𝟙′[𝑩,+∞[ (𝒚)]
𝒇𝒀 (𝒚) = 𝜹𝑩 (𝒚) + 𝒇𝑿 (𝒚) − 𝒇𝑿 (𝒚)𝟙[𝑩,+∞[ (𝒚) − 𝑭𝑿 (𝒚)𝜹𝑩 (𝒚)
◦ 𝑺𝒊 𝒚 < 𝟏 ⟹ 𝑭𝑿 (𝒚) = 𝟎, 𝒇𝑿 (𝒚) = 𝟎, 𝜹𝑩 (𝒚) = 𝟎 𝒆𝒕 𝟙[𝑩,+∞[ (𝒚) = 𝟎, 𝒅𝒐𝒏𝒄, 𝒇𝒀 (𝒚) = 𝟎
𝟏 𝟒 𝟒
◦ 𝑺𝒊 𝟏 ≤ 𝒚 < 𝑩 ⟹ 𝑭𝑿 (𝒚) = 𝟏 − 𝟒 , 𝒇𝑿 (𝒚) = 𝟓 , 𝜹𝑩 (𝒚) = 𝟎 𝒆𝒕 𝟙[𝑩,+∞[ (𝒚) = 𝟎, 𝒅𝒐𝒏𝒄, 𝒇𝒀 (𝒚) = 𝟓
𝒚 𝒚 𝒚
𝟏
◦ 𝑺𝒊 𝒚 = 𝑩 ⟹ 𝑭𝑿 (𝒚) = 𝟏 − 𝟒 , 𝒇𝑿 (𝒚) = 𝟎, 𝜹𝑩 (𝒚) = 𝟏 𝒆𝒕 𝟙[𝑩,+∞[ (𝒚) = 𝟏,
𝑩
𝟏 𝟏
𝒅𝒐𝒏𝒄, 𝑷(𝒀 = 𝒚) = 𝟏 − (𝟏 − 𝟒 ) = 𝟒
𝑩 𝑩
◦ 𝑺𝒊 𝒚 > 𝑩 ⟹ 𝒇𝒀 (𝒚) =𝟎
𝑶𝒏 𝒗é𝒓𝒊𝒇𝒊𝒆 𝒃𝒊𝒆𝒏 𝒒𝒖𝒆 𝒀 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒎𝒊𝒙𝒕𝒆 ∶

331 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝑩
𝟏 𝟏 𝟏
∫ 𝒇𝒀 (𝒚)𝒅𝒚 + 𝑷(𝒀 = 𝑩) = 𝑭𝒀 (𝑩) − 𝑭𝒀 (𝟏) + 𝟒 = 𝟏 − 𝟒 − 𝟎 + 𝟒 = 𝟏
𝟏 𝑩 𝑩 𝑩
𝟒
𝒇𝑿 (𝒚) = 𝟓 , 𝒔𝒊 𝟏 ≤ 𝒚 < 𝑩
𝒚
𝑫’𝒐ù 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒎𝒊𝒙𝒕𝒆 𝒅𝒆 𝒀 : 𝒇𝒀 (𝒚) = 𝟏
𝑷(𝒀 = 𝑩) = 𝟏 − 𝑭𝒀 (𝑩) = 𝟒 , 𝒔𝒊 𝒚 = 𝑩
𝑩
{𝟎 𝒂𝒊𝒍𝒍𝒆𝒖𝒓𝒔
7)
▪𝑪𝒂𝒍𝒄𝒖𝒍𝒐𝒏𝒔 𝒅’𝒂𝒃𝒐𝒓𝒅, 𝑬(𝒀):
𝑩
𝟒 𝑩
𝟏 𝟏 𝑩 𝟏 𝟒 𝟏 𝑩 𝟏
𝑬(𝒀) = (∫ 𝒚𝒇𝒀 (𝒚)𝒅𝒚) + 𝑩𝑷(𝒀 = 𝑩) = (∫ 𝟒
𝒅𝒚) + = 𝟒 [ ] + = − [ ] +
𝟏 𝟏 𝒚 𝑩𝟑 −𝟑𝒚𝟑 𝟏 𝑩𝟑 𝟑 𝒚𝟑 𝟏 𝑩𝟑
𝟒 𝟏 𝟏 𝟒 𝟒 𝟑 𝟒 𝟏
𝑬(𝒀) = − ( 𝟑 − 𝟏) + 𝟑 = − + = −
𝟑 𝑩 𝑩 𝟑 𝟑𝑩𝟑 𝟑𝑩𝟑 𝟑 𝟑𝑩𝟑
𝟒 𝟏
𝑬(𝒀) = −
𝟑 𝟑𝑩𝟑
𝟕 𝟒 𝟏 𝟕 𝟏 𝟏 𝟏 𝟏
▪𝑬(𝒀) = ⟺ − 𝟑
= ⟺ 𝟑
= ⟺ 𝟑 = ⟺ 𝑩𝟑 = 𝟐
𝟔 𝟑 𝟑𝑩 𝟔 𝟑𝑩 𝟔 𝑩 𝟐
𝟕 𝟑
𝑬(𝒀) = ⟺ 𝑩 = √𝟐 ≅ 𝟏, 𝟐𝟔
𝟔
𝑬𝒏 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒍𝒆 𝒓𝒆𝒎𝒃𝒐𝒖𝒓𝒔𝒆𝒎𝒆𝒏𝒕 𝒆𝒔𝒕 𝒍é𝒈è𝒓𝒆𝒎𝒆𝒏𝒕 𝒊𝒏𝒇𝒆𝒓𝒊𝒆𝒖𝒓𝒆 à 𝒍𝒂 𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒅𝒆𝒔 𝒑𝒆𝒓𝒕𝒆𝒔

Exercice 9 : (Indépendance de variables aléatoires gaussiennes et corrélation)

ÉNONCÉ
𝑺𝒐𝒊𝒆𝒏𝒕 𝑿 𝒆𝒕 𝒀 𝒅𝒆𝒖𝒙 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒎é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒔𝒖𝒊𝒗𝒂𝒏𝒕 𝒍𝒂 𝒎ê𝒎𝒆 𝒍𝒐𝒊 𝓝(𝟎, 𝟏) .
𝑶𝒏 𝒑𝒐𝒔𝒆 𝑼 = 𝑿 − 𝒀 𝒆𝒕 𝑽 = 𝑿 + 𝒀 .
𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝝆 𝒍𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝒅𝒆 𝒄𝒐𝒓𝒓é𝒍𝒂𝒕𝒊𝒐𝒏 𝒍𝒊𝒏é𝒂𝒊𝒓𝒆 𝒆𝒏𝒕𝒓𝒆 𝑼 𝒆𝒕 𝑽 .

Corrigé
𝑿, 𝒀 ↝ 𝓝(𝟎, 𝟏) 𝑼 = 𝑿 − 𝒀 ↝ 𝓝(𝑬(𝑿 − 𝒀), 𝑽(𝑿 − 𝒀))
{ ⇒{
𝑿 𝒆𝒕 𝒀 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝑽 = 𝑿 + 𝒀 ↝ 𝓝(𝑬(𝑿 + 𝒀), 𝑽(𝑿 + 𝒀))

𝒐𝒓, 𝑬(𝑼) = 𝑬(𝑿 − 𝒀) = 𝑬(𝑿) − 𝑬(𝒀) = 𝟎 , 𝑬(𝑽) = 𝑬(𝑿 + 𝒀) = 𝑬(𝑿) + 𝑬(𝒀) = 𝟎

𝒆𝒕 𝑽(𝑼) = 𝑽(𝑿 − 𝒀) = 𝑽(𝑿 + 𝒀) = 𝑽(𝑽) = 𝑽(𝑿) + 𝑽(𝒀) = 𝟐
𝒂𝒊𝒏𝒔𝒊 , 𝑼 𝒆𝒕 𝑽 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒅𝒆 𝒎ê𝒎𝒆 𝒍𝒐𝒊 𝓝(𝟎, 𝟐)
𝑪𝒂𝒍𝒄𝒖𝒍𝒐𝒏𝒔 𝑪𝒐𝒗(𝑼, 𝑽) ∶
𝑪𝒐𝒗(𝑼, 𝑽) = 𝑪𝒐𝒗[(𝑿 − 𝒀), (𝑿 + 𝒀)] = 𝑪𝒐𝒗(𝑿, 𝑿) + ⏟
𝑪𝒐𝒗(𝑿, 𝒀) − ⏟
𝑪𝒐𝒗(𝒀, 𝑿) − 𝑪𝒐𝒗(𝒀, 𝒀)
𝟎 𝟎

= 𝑽(𝑿) − 𝑽(𝒀)

𝑪𝒐𝒗(𝑼, 𝑽) = 𝟎 é𝒒𝒖𝒊𝒗𝒂𝒖𝒕 à 𝒅𝒊𝒓𝒆 𝒒𝒖𝒆 𝑿 𝒆𝒕 𝒀 𝒏𝒐𝒏 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒏𝒐𝒏 𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔 𝒎𝒂𝒊𝒔 𝒂𝒖𝒔𝒔𝒊

332 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔, 𝒑𝒖𝒊𝒔𝒒𝒖’𝒊𝒍 𝒔’𝒂𝒈𝒊𝒕 𝒅𝒆 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒅𝒆 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒏𝒐𝒓𝒎𝒂𝒍𝒆

𝑪𝒐𝒗(𝑼, 𝑽)
𝝆= = 𝟎 𝑼 𝒆𝒕 𝑽 𝒔𝒐𝒏𝒕 𝒂𝒖𝒔𝒔𝒊 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔
√𝑽(𝑼)𝑽(𝑽)

Exercice 10 : (Transformations de variables-Inégalité de Jensen)

ÉNONCÉ
𝑳𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒇(𝒙) 𝒅′ 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝑿 𝒔′ é𝒄𝒓𝒊𝒕 𝒔𝒐𝒖𝒔 𝒍𝒂 𝒇𝒐𝒓𝒎𝒆 ∶
𝟐
𝒇𝑿 (𝒙) = { 𝒙𝟑 , 𝒑𝒐𝒖𝒓 𝒙 ≥ 𝟏
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
1) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝑿
2) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒂 𝒎é𝒅𝒊𝒂𝒏𝒆 𝒅𝒆 𝑿
𝟏
3) 𝑶𝒏 𝒅é𝒇𝒊𝒏𝒊𝒕, 𝒁 = 𝑿 . 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒅𝒆 𝒁
𝟏 𝟏
4) 𝑪𝒐𝒎𝒑𝒂𝒓𝒆𝒓 𝑬 (𝑿) à 𝑬(𝑿)

Corrigé
𝒙
𝒙 𝒙 𝒕−𝟑+𝟏 𝟏 𝒙 𝟏
1) ∀𝒙 ≥ 𝟏, 𝑭𝑿 (𝒙) = 𝑷(𝑿 ≤ 𝒙) = ∫𝟏 𝒇𝑿 (𝒕)𝒅𝒕 = 𝟐 ∫𝟏 𝒕−𝟑 𝒅𝒕 = 𝟐 [−𝟑+𝟏] = − [𝒕𝟐 ] = 𝟏 − 𝒙𝟐
𝟏 𝟏

𝟎 , 𝒔𝒊 𝒙 < 𝟏
′ 𝟏
𝑫 𝒐ù, 𝑭𝑿 (𝒙) = {
𝟏 − 𝟐 , 𝒔𝒊 𝒙 ≥ 𝟏
𝒙
𝟏 𝟏 𝟏 𝟏 𝟏
2) 𝑭𝑿 (𝑴𝒆 ) = 𝟐 ⟺ 𝟏 − 𝑴𝟐 = 𝟐 ⟺ 𝑴𝟐 = 𝟐 ⟺ 𝑴𝟐𝒆 = 𝟐, 𝑫′ 𝒐ù 𝑴𝒆 = √𝟐
𝒆 𝒆
𝑿(𝛀) = [𝟏, +∞[
3) { 𝟏 ⇒ 𝒁(𝛀) = ]𝟎, 𝟏]
𝒁=𝑿

𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒁 ; ∀𝒛 ∈ ]𝟎, 𝟏] 𝒆𝒕 ∀𝒙 ∈ [𝟏, +∞[ ∶

𝟏 𝟏 𝟏 𝟏
𝑭𝒁 (𝒛) = 𝑷(𝒁 ≤ 𝒛) = 𝑷 ( ≤ 𝒛) = 𝑷 (𝑿 ≥ ) = 𝟏 − 𝑷 (𝑿 < ) = 𝟏 − 𝑭𝑿 ( )
𝑿 𝒛 𝒛 𝒛
𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒁 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏 ∶
𝟏
𝒅𝑭𝒁 (𝒛) 𝒅 [𝟏 − 𝑭𝑿 ( 𝒛 )] 𝟏 ′ 𝟏 𝟏 𝟐
∀𝒛 ∈ ]𝟎, 𝟏] ; 𝒇𝒁 (𝒛) = = = − ( ) 𝒇𝑿 ( ) = 𝟐 × = 𝟐𝒛
𝒅𝒛 𝒅𝒛 𝒛 𝒛 𝒛 𝟏 𝟑
(𝒛)

𝟐𝒛 , 𝒔𝒊 𝒛 ∈ ]𝟎, 𝟏]
𝑫′ 𝒐ù, 𝒇𝒁 (𝒛) = {
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
𝟏
𝟏 𝟏 𝟏 𝒛𝟐+𝟏 𝟐 𝟐
4) 𝑬 (𝑿) = 𝑬(𝒁) = ∫𝟎 𝒛𝒇𝒁 (𝒛)𝒅𝒛 = 𝟐 ∫𝟎 𝒛𝟐 𝒅𝒛 = 𝟐 [ 𝟐+𝟏 ] = 𝟑 [𝒛𝟑 ]𝟏𝟎 = 𝟑
𝟎

333 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
+∞ +∞
𝒅𝒙
𝑬(𝑿) = ∫ 𝒙𝒇𝑿 (𝒙)𝒅𝒙 = 𝟐 ∫ , 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 (𝒊𝒏𝒕é𝒈𝒓𝒂𝒍𝒆 𝒅𝒆 𝑹𝒊𝒆𝒎𝒂𝒏𝒏, 𝒑𝒐𝒖𝒓 𝜶 = 𝟐 > 𝟏)
𝟏 𝟏 𝒙𝟐
+∞
𝒙−𝟐+𝟏 𝟏 +∞ 𝟏
𝑬(𝑿) = 𝟐 [ ] = −𝟐 [ ] = −𝟐 [( 𝐥𝐢𝐦 ) − 𝟏] = 𝟐
−𝟐 + 𝟏 𝟏 𝒙𝟏 𝒙→+∞ 𝒙

𝟏 𝟏
𝑫′ 𝒐ù, < 𝑬( )
𝑬(𝑿) 𝑿

𝑶𝒏 𝒑𝒐𝒖𝒓𝒓𝒂 𝒂𝒓𝒓𝒊𝒗𝒆𝒓 à 𝒄𝒆 𝒓é𝒔𝒖𝒍𝒕𝒂𝒕 𝒂𝒖𝒕𝒓𝒆𝒎𝒆𝒏𝒕, 𝒆𝒏 𝒖𝒕𝒊𝒍𝒊𝒔𝒂𝒏𝒕 𝒍′𝒊𝒏é𝒈𝒂𝒍𝒊𝒕é 𝒅𝒆 𝑱𝒆𝒏𝒔𝒆𝒏 ∶ 𝑰𝒍 𝒔𝒖𝒇𝒇𝒊𝒕

𝟏 𝟏
𝒅𝒆 𝒓𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒓 𝒒𝒖𝒆 𝝋(𝒖) = 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒄𝒐𝒏𝒗𝒆𝒙𝒆 𝒔𝒖𝒓 [𝟏, +∞[ , 𝒑𝒖𝒊𝒔𝒒𝒖𝒆 𝝋′ (𝒖) = − 𝟐 ⇒
𝒖 𝒖
𝟐
𝝋′′ (𝒖) = > 𝟎 , ∀𝒖 ∈ [𝟏, +∞[ . 𝑫′ 𝒂𝒑𝒓è𝒔 𝒍′ 𝒊𝒏é𝒈𝒂𝒍𝒊𝒕é 𝒅𝒆 𝑱𝒆𝒏𝒔𝒆𝒏, 𝒐𝒏 𝒓𝒆𝒕𝒓𝒐𝒖𝒗𝒆 ∶
𝒖𝟑
𝟏 𝟏
𝝋[𝑬(𝑿)] ≤ 𝑬[𝝋(𝑿)], 𝒂𝒖𝒕𝒓𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒕 ∶ < 𝑬( )
𝑬(𝑿) 𝑿

Exercice 11 : (Corrélation entre variables la loi normale centrée réduite)

ÉNONCÉ
𝑶𝒏 𝒄𝒐𝒏𝒔𝒊𝒅è𝒓𝒆 𝒅𝒆𝒖𝒙 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝑿 𝒆𝒕 𝒀 𝒏𝒐𝒓𝒎𝒂𝒍𝒆𝒔 𝒄𝒆𝒏𝒕𝒓é𝒆𝒔 𝒓é𝒅𝒖𝒊𝒕𝒆𝒔 𝒂𝒗𝒆𝒄 𝒖𝒏𝒆 𝒄𝒐𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆,
𝑪𝒐𝒗(𝑿, 𝒀) = 𝒄 . 𝑶𝒏 𝒑𝒐𝒔𝒆 ∶ 𝒁 = 𝑿 − 𝒂𝒀
1) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒂 𝒑𝒐𝒖𝒓 𝒒𝒖𝒆 𝑿 𝒆𝒕 𝒁 𝒔𝒐𝒊𝒆𝒏𝒕 𝒏𝒐𝒏 𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔
2) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓, 𝒂𝒍𝒐𝒓𝒔 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒅𝒆 𝒁

Corrigé
1) 𝑿, 𝒀 ↝ 𝓝(𝟎, 𝟏) ⇒ 𝑬(𝑿) = 𝑬(𝒀) = 𝟎 𝒆𝒕 𝑽(𝑿) = 𝑽(𝒀) = 𝟏
𝑿 𝒆𝒕 𝒁 𝒔𝒐𝒊𝒆𝒏𝒕 𝒏𝒐𝒏 𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔 ⟺ 𝑪𝒐𝒗(𝑿, 𝒁) = 𝟎 ⟺ 𝑪𝒐𝒗[𝑿, (𝑿 − 𝒂𝒀)] = 𝟎
𝟏
⟺ 𝑪𝒐𝒗(𝑿, 𝑿) − 𝒂𝑪𝒐𝒗(𝑿, 𝒀) = 𝟎 ⟺ 𝑽(𝑿) − 𝒂𝑪𝒐𝒗(𝑿, 𝒀) = 𝟎 ⟺ 𝟏 − 𝒂𝒄 = 𝟎 ⟺ 𝒂 = ,𝒄 ≠ 𝟎
𝒄
𝑿 𝒆𝒕 𝒀 𝒔𝒐𝒊𝒆𝒏𝒕 𝒏𝒐𝒏 𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔 (𝒄. −à − 𝒅. 𝒄 ≠ 𝟎)
′ 𝟏
𝑫 𝒐ù, 𝑿 𝒆𝒕 𝒁 𝒔𝒐𝒊𝒆𝒏𝒕 𝒏𝒐𝒏 𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔 ⟺ {
𝒂=
𝒄
𝟏
2) 𝑽(𝒁) = 𝑽(𝑿 − 𝒂𝒀) = 𝑽(𝑿) + 𝒂𝟐 𝑽(𝒀) − 𝟐𝒂𝑪𝒐𝒗(𝑿, 𝒀) = 𝟏 + 𝒂𝟐 − 𝟐𝒂𝒄 = 𝟏 + 𝒄𝟐 − 𝟐

𝟏 𝟏 − 𝒄𝟐
𝑫′ 𝒐ù, 𝒑𝒐𝒖𝒓 𝒄 ≠ 𝟎, 𝒆𝒕 𝒂 = , 𝑽(𝒁) =
𝒄 𝒄𝟐

334 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
Exercice 12 : (Transformations de variables-Loi Exponentielle)
ÉNONCÉ
𝑺𝒐𝒊𝒕 𝑿 ↝ 𝓔(𝟏), 𝒄𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒂 𝒅. 𝒅. 𝒑. 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 = 𝐥𝐧 𝑿

Corrigé
𝒆−𝒙 , 𝒑𝒐𝒖𝒓 𝒙 ≥ 𝟎
𝑿 ↝ 𝓔(𝟏) ⇒ 𝒇𝑿 (𝒙) = {
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
𝑿(𝛀) = ℝ+
{ ⇒ 𝒀(𝛀) = ℝ , 𝒑𝒐𝒖𝒓 𝒙 > 𝟎
𝒀 = 𝐥𝐧 𝑿
𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 :
∀𝒚 ∈ ℝ 𝒆𝒕 ∀𝒙 ∈ ]𝟎, +∞[ , 𝑭𝒀 (𝒚) = 𝑷(𝒀 ≤ 𝒚) = 𝑷(𝐥𝐧 𝑿 ≤ 𝒚) = 𝑷(𝑿 ≤ 𝒆𝒚 ) = 𝑭𝑿 (𝒆𝒚 )
𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏 ∶
𝒅𝑭𝒀 (𝒚) 𝒅𝑭𝑿 (𝒆𝒚 ) 𝒚 𝒚
∀𝒚 ∈ ℝ ; 𝒇𝒀 (𝒚) = = = (𝒆𝒚 )′ 𝒇𝑿 (𝒆𝒚 ) = 𝒆𝒚 𝒆−(𝒆 ) = 𝒆(𝒚−𝒆 )
𝒅𝒚 𝒅𝒚
𝒚
𝑫′ 𝒐ù 𝒇𝒀 (𝒚) = 𝒆(𝒚−𝒆 ) , ∀𝒚 ∈ ℝ

Exercice 13 : (Transformations de variables-Loi Normale centrée-réduite)

ÉNONCÉ
𝑺𝒐𝒊𝒕 𝑿 ↝ 𝓝(𝟎, 𝟏), 𝒄𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒂 𝒅. 𝒅. 𝒑. 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 = 𝒆𝑿

Corrigé
𝟏 𝒙𝟐
−
𝑿 ↝ 𝓝(𝟎, 𝟏) ⇒ ∀𝒙 ∈ ℝ , 𝒇𝑿 (𝒙) = 𝒆 𝟐
√𝟐𝝅
𝑿(𝛀) = ℝ
{ ⇒ 𝒀(𝛀) = ]𝟎, +∞[
𝒀 = 𝒆𝑿
𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 ; ∀𝒙 ∈ ℝ 𝒆𝒕 ∀𝒚 ∈ ]𝟎, +∞[ :
𝑭𝒀 (𝒚) = 𝑷(𝒀 ≤ 𝒚) = 𝑷(𝒆𝑿 ≤ 𝒚) = 𝑷(𝑿 ≤ 𝐥𝐧 𝒚) = 𝑭𝑿 (𝐥𝐧 𝒚)
𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏 ∶
𝒅𝑭𝒀 (𝒚) 𝒅𝑭𝑿 (𝐥𝐧 𝒚) 𝟏 𝟏 −(𝐥𝐧 𝒚)𝟐
∀𝒚 ∈ ]𝟎, +∞[ ; 𝒇𝒀 (𝒚) = = = (𝐥𝐧 𝒚)′ 𝒇𝑿 (𝐥𝐧 𝒚) = × 𝒆 𝟐
𝒅𝒚 𝒅𝒚 𝒚 √𝟐𝝅

𝟏 (𝐥𝐧 𝒚)𝟐
−
𝒆 𝟐 , 𝒔𝒊 𝒚 > 𝟎
𝑫′ 𝒐ù 𝒇𝒀 (𝒚) = {√𝟐𝝅𝒚𝟐
𝟎 , 𝒔𝒊 𝒏𝒐𝒏

335 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
Exercice 14 : (Transformations de variables- Loi khi‑deux à 1 degré de liberté)
ÉNONCÉ
𝑺𝒐𝒊𝒕 𝑿 ↝ 𝓝(𝟎, 𝟏) 𝒆𝒕 𝒅é𝒇𝒊𝒏𝒊𝒔𝒔𝒐𝒏𝒔 𝒀 = 𝑿𝟐 .
1) 𝑻𝒓𝒐𝒖𝒗𝒆𝒓 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒀 𝒆𝒕 𝒔𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é.
𝑵. 𝑩. ∶ 𝑪𝒆𝒕𝒕𝒆 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏 𝒆𝒔𝒕 𝒂𝒑𝒑𝒆𝒍é𝒆 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒅𝒆 𝒌𝒉𝒊‑𝒅𝒆𝒖𝒙 à 𝟏 𝒅𝒆𝒈𝒓é 𝒅𝒆 𝒍𝒊𝒃𝒆𝒓𝒕é ∶ 𝝌𝟐 (𝟏)
2) 𝑺𝒐𝒊𝒕 𝒎𝒂𝒊𝒏𝒕𝒆𝒏𝒂𝒏𝒕 𝒍𝒂 𝒗. 𝒂. 𝑼 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒑𝒂𝒓 : 𝑼 = 𝐥𝐧(𝒀) . 𝑻𝒓𝒐𝒖𝒗𝒆𝒓 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆
𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝑼 𝒆𝒕 𝒔𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é.

Corrigé
𝒙𝟐
𝟏 𝑿(𝛀) = ℝ
1) 𝑶𝒏 𝒂 ∶ 𝑿 ↝ 𝓝(𝟎, 𝟏) ⇒ ∀𝒙 ∈ ℝ , 𝒇𝑿 (𝒙) = 𝒆− 𝟐 ; { ⇒ 𝒀(𝛀) = [𝟎, +∞[
√𝟐𝝅 𝒀 = 𝑿𝟐
𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 ; ∀(𝒙, 𝒚) ∈ ℝ × [𝟎, +∞[ ∶

𝑭𝒀 (𝒚) = 𝑷(𝒀 ≤ 𝒚) = 𝑷(𝑿𝟐 ≤ 𝒚) = 𝑷(−√𝒚 ≤ 𝑿 ≤ √𝒚) = 𝑭𝑿 (√𝒚) − 𝑭𝑿 (−√𝒚) = 𝚽(√𝒚) − 𝚽(−√𝒚)

= 𝚽(√𝒚) − (𝟏 − 𝚽(√𝒚)) = 𝟐𝚽(√𝒚) − 𝟏

𝒅[𝟐𝚽(√𝒚) − 𝟏]
𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏: ∀𝒚 ∈ ]𝟎, +∞[ ; 𝒇𝒀 (𝒚) =
𝒅𝒚
𝟐
′ 𝟏 𝟏 (√𝒚) 𝟏 𝒚
−
∀𝒚 ∈ ]𝟎, +∞[ ; 𝒇𝒀 (𝒚) = 𝟐(√𝒚) 𝒇𝑿 (√𝒚) = 𝒇𝑿 (√𝒚) = 𝒆 𝟐 = 𝒆− 𝟐
√𝒚 √𝟐𝝅√𝒚 √𝟐𝝅𝒚

𝑪𝒐𝒎𝒎𝒆 𝑿 ↝ 𝓝(𝟎, 𝟏) ⇒ 𝒀 = 𝑿𝟐 ↝ 𝝌𝟐 (𝟏) , 𝒐𝒏 𝒆𝒏 𝒅é𝒅𝒖𝒊𝒕 𝒂𝒍𝒐𝒓𝒔 𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝝌𝟐 (𝟏) ∶

𝟏 𝒚
𝒆−𝟐 , 𝒔𝒊𝒚 ∈ ]𝟎, +∞[
𝒇𝒀 (𝒚) = {√𝟐𝝅𝒚
𝟎 , 𝒔𝒊 𝒏𝒐𝒏

𝟏 𝒚 𝒀(𝛀) = ]𝟎, +∞[

2) 𝑶𝒏 𝒂 ∶ 𝒀 ↝ 𝝌𝟐 (𝟏) ⇒ ∀𝒚 ∈ ]𝟎, +∞[ , 𝒇𝒀 (𝒚) = 𝒆− 𝟐 ; { ⇒ 𝑼(𝛀) = ℝ
√𝟐𝝅𝒚 𝑼 = 𝐥𝐧(𝒀)
𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝑼 ; ∀(𝒚, 𝒖) ∈ ]𝟎, +∞[ × ℝ ∶
𝒖
𝑭𝑼 (𝒖) = 𝑷(𝑼 ≤ 𝒖) = 𝑷(𝐥𝐧(𝒀) ≤ 𝒖) = 𝑷(𝒀 ≤ 𝒆𝒖 ) = 𝑭𝒀 (𝒆𝒖 ) = 𝟐𝚽(√𝒆𝒖 ) − 𝟏 = 𝟐𝚽 (𝒆𝟐 ) − 𝟏
𝒖
𝒅 [𝟐𝚽 (𝒆𝟐 ) − 𝟏]
𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝑼 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏 ∶ ∀𝒖 ∈ ℝ ; 𝒇𝑼 (𝒖) =
𝒅𝒖
𝒖 𝟐
𝒖
𝒖 ′ 𝒖 𝒖 𝟏 −(𝒆𝟐 ) ⁄𝟐 𝒆𝟐 (𝒆𝒖 ) 𝟏 (𝒖−𝒆𝒖 )
−
∀𝒖 ∈ ℝ ; 𝒇𝑼 (𝒖) = 𝟐 (𝒆𝟐 ) 𝒇𝑿 (𝒆𝟐 ) = 𝒆𝟐 [ 𝒆 ]= 𝒆 𝟐 = 𝒆 𝟐
√𝟐𝝅 √𝟐𝝅 √𝟐𝝅

𝟏 (𝒖−𝒆𝒖 )
∀𝒖 ∈ ℝ ; 𝒇𝑼 (𝒖) = 𝒆 𝟐
√𝟐𝝅

336 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
Exercice 15 : (Transformations de variables-Loi Uniforme continue-Loi Exponentielle)
ÉNONCÉ
𝟏
𝑺𝒐𝒊𝒕 𝑽 ↝ 𝓤[𝟎,𝟏] . 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝑾 = − 𝐥𝐧(𝑽) 𝒆𝒕 𝒔𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é. 𝝀 é𝒕𝒂𝒏𝒕
𝝀
𝒖𝒏 𝒓é𝒆𝒍 𝒔𝒕𝒓𝒊𝒄𝒕𝒆𝒎𝒆𝒏𝒕 𝒑𝒐𝒔𝒊𝒕𝒊𝒇 .
𝑫𝒆 𝒒𝒖𝒆𝒍𝒍𝒆 𝒍𝒐𝒊 𝒔’𝒂𝒈𝒊𝒕‑𝒊𝒍 ?

Corrigé
𝟏 , 𝒔𝒊 𝒗 ∈ [𝟎, 𝟏]
𝑶𝒏 𝒂 𝑽 ↝ 𝓤[𝟎,𝟏] ⇒ ∀𝒗 ∈ ℝ , 𝒇𝑽 (𝒗) = {
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
𝟏
∀𝒗 ∈ [𝟎, 𝟏], 𝒐𝒏 𝒂 ∶ 𝐥𝐧(𝒗) ≤ 𝟎 , 𝒐𝒓 𝝀 > 𝟎 , 𝒅𝒐𝒏𝒄 − 𝐥𝐧(𝑽) ≥ 𝟎
𝝀
𝑽(𝛀) = [𝟎, 𝟏]
{ 𝟏 ⇒ 𝑾(𝛀) = [𝟎, +∞[
𝑾 = − 𝐥𝐧(𝑽)
𝝀
𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝑾 ; ∀𝒗 ∈ [𝟎, 𝟏] 𝒆𝒕 ∀𝒘 ∈ [𝟎, +∞[ ∶
𝟏
𝑭𝑾 (𝒘) = 𝑷(𝑾 ≤ 𝒘) = 𝑷 (− 𝐥𝐧(𝑽) ≤ 𝒘) = 𝑷(𝐥𝐧(𝑽) ≥ −𝝀𝒘) = 𝑷(𝑽 ≥ 𝒆−𝝀𝒘 ) = 𝟏 − 𝑷(𝑽 < 𝒆−𝝀𝒘 )
𝝀
= 𝟏 − 𝑷(𝑽 ≤ 𝒆−𝝀𝒘 ) = 𝟏 − 𝑭𝑽 (𝒆−𝝀𝒘 )

𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝑾 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏 ∶

𝒅[𝟏 − 𝑭𝑽 (𝒆−𝝀𝒘 )] ′
∀𝒘 ∈ [𝟎, +∞[ ; 𝒇𝑾 (𝒘) = = −(𝒆−𝝀𝒘 ) 𝒇𝑽 (𝒆−𝝀𝒘 ) = 𝝀𝒆−𝝀𝒘 𝒇𝑽 (𝒆−𝝀𝒘 )
𝒅𝒘
𝑶𝒓 𝒘 ∈ [𝟎, +∞[ ⟹ −𝝀𝒘 ≤ 𝟎 ⟹ 𝟎 ≤ 𝒆−𝝀𝒘 ≤ 𝟏 𝒆𝒕 𝒇𝑽 (𝒆−𝝀𝒘 ) = 𝟏

𝝀𝒆−𝝀𝒘 , 𝒔𝒊 𝒘 ∈ [𝟎, +∞[

𝑫’𝒐ù 𝒍𝒂 𝒅. 𝒅. 𝒑. 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝑾 ∶ 𝒇𝑾 (𝒘) = { 𝒆𝒕 𝑾 ↝ 𝓔(𝝀)
𝟎 , 𝒔𝒊 𝒏𝒐𝒏

Exercice 16 : (Transformations de variables-Loi de Weibull-Loi Exponentielle)

ÉNONCÉ
𝑺𝒐𝒊𝒕 𝑿 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖é𝒆 𝒔𝒖𝒊𝒗𝒂𝒏𝒕 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝑾𝒆𝒊𝒃𝒖𝒍𝒍 𝒅𝒆 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆𝒔 𝜶 > 𝟎 𝒆𝒕
𝜷 > 𝟎 . 𝑳𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝑿 𝒆𝒔𝒕 𝒅𝒐𝒏𝒄 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒑𝒂𝒓 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 ∶
𝜷 𝒙 𝜷−𝟏 −( 𝒙 )𝜷
𝒇𝑿 (𝒙) = { 𝜶) (𝜶)
( 𝒆 𝜶 , 𝒑𝒐𝒖𝒓 𝒙 ≥ 𝟎
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
𝑺𝒐𝒊𝒕 𝒀 = (𝑿⁄𝜶)𝜷 . 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒀. 𝑫𝒆 𝒒𝒖𝒆𝒍𝒍𝒆 𝒍𝒐𝒊 𝒔’𝒂𝒈𝒊𝒕‑𝒊𝒍 ?

337 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
Corrigé
∀𝒙 ∈ [𝟎, +∞[, 𝒐𝒏 𝒂 ∶ (𝒙⁄𝜶)𝜷 > 𝟎 , 𝒑𝒖𝒊𝒔𝒒𝒖𝒆 𝜶 > 𝟎 𝒆𝒕 𝜷 > 𝟎
𝑿(𝛀) = [𝟎, +∞[
{ ⇒ 𝒀(𝛀) = [𝟎, +∞[
𝒀 = (𝑿⁄𝜶)𝜷
𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 :

𝟐
𝑿 𝜷 𝑿 𝟏 𝟏 𝟏
∀(𝒙, 𝒚) ∈ [𝟎, +∞[ , 𝑭𝒀 (𝒚) = 𝑷(𝒀 ≤ 𝒚) = 𝑷 (( ) ≤ 𝒚) = 𝑷 ( ≤ 𝒚𝜷 ) = 𝑷 (𝑿 ≤ 𝜶𝒚𝜷 ) = 𝑭𝑿 (𝜶𝒚𝜷 )
𝜶 𝜶

𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏 ∶

𝟏
𝒅 [𝑭𝑿 (𝜶𝒚𝜷 )] 𝟏 ′ 𝟏
∀𝒘 ∈ [𝟎, +∞[ ; 𝒇𝒀 (𝒚) = = 𝜶 (𝒚𝜷 ) 𝒇𝑿 (𝜶𝒚𝜷 )
𝒅𝒚
𝟏 𝜷
𝜷−𝟏 𝜷
(𝜶𝒚 )
𝟏
− 𝜷
(𝜶𝒚𝜷 ) 𝜶 𝟏
𝟏 𝜷−𝟏 −(𝒚𝜷 )
𝜶 ( 𝟏−𝟏) 𝜷 𝟏
( −𝟏)
= [ 𝒚𝜷 ] ( ) 𝒆 ( ) = [𝒚 𝜷 ] [(𝒚𝜷 ) 𝒆 ]
𝜷 𝜶 𝜶
( )
[ ]
𝟏 𝟏
( −𝟏) (𝟏− ) −𝒚
𝒚𝜷
=⏟ 𝒚 𝜷 𝒆
𝟏

𝒆−𝒚 , 𝒔𝒊 𝒚 ∈ [𝟎, +∞[

𝑫’𝒐ù 𝒍𝒂 𝒅. 𝒅. 𝒑. 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝑾 ∶ 𝒇𝒀 (𝒚) = { 𝒆𝒕 𝒀 ↝ 𝓔(𝟏)
𝟎 , 𝒔𝒊 𝒏𝒐𝒏

Exercice 17 : (Transformations de variables-Loi de Cauchy)

ÉNONCÉ
𝑺𝒐𝒊𝒕 𝑽 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖é𝒆 𝒔𝒖𝒊𝒗𝒂𝒏𝒕 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝑪𝒂𝒖𝒄𝒉𝒚 𝓒(𝟎 , 𝟏)𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é ∶
𝟏⁄𝝅
∀𝒗 ∈ ℝ , 𝒇𝑽 (𝒗) = .
𝟏 + 𝒗𝟐
𝟏
𝑺𝒐𝒊𝒕 𝒁 = . 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒁. 𝑸𝒖𝒆 𝒓𝒆𝒎𝒂𝒓𝒒𝒖𝒆𝒛‑𝒗𝒐𝒖𝒔 ?
𝑽
Corrigé
𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒁 :
𝟏 𝟏 𝟏 𝟏
∀(𝒗, 𝒛) ∈ (ℝ∗ )𝟐 , 𝑭𝒁 (𝒛) = 𝑷(𝒁 ≤ 𝒛) = 𝑷 ( ≤ 𝒛) = 𝑷 (𝑽 ≥ ) = 𝟏 − 𝑷 (𝑽 ≤ ) = 𝟏 − 𝑭𝑽 ( )
𝑽 𝒛 𝒛 𝒛
𝒍𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒁 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏 ∶

338 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝟏
𝒅 [𝟏 − 𝑭𝑽 ( 𝒛 )] 𝟏 ′ 𝟏 𝟏 𝟏⁄𝝅 𝟏 𝟏⁄𝝅 𝟏⁄𝝅
∀𝒛 ∈ ℝ∗ ; 𝒇𝒁 (𝒛) = = − ( ) 𝒇𝑽 ( ) = 𝟐 [ 𝟐 ] = 𝟐[ 𝟐 ]=
𝒅𝒛 𝒛 𝒛 𝒛 𝟏 𝒛 𝟏+𝒛 𝟏 + 𝒛𝟐
𝟏 + (𝒛) 𝒛𝟐
𝑳𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒁 𝒆𝒔𝒕 𝒂𝒖𝒔𝒔𝒊 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒔𝒖𝒊𝒗𝒂𝒏𝒕 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒅𝒆 𝑪𝒂𝒖𝒄𝒉𝒚 ∶

𝑽 ↝ 𝓒(𝟎 , 𝟏)
{ 𝟏 ⇒ 𝒁 ↝ 𝓒(𝟎 , 𝟏)
𝒁=
𝑽

Exercice 18 : (Répartition des richesses-Loi de Pareto)

ÉNONCÉ
𝑪𝒐𝒏𝒔𝒊𝒅é𝒓𝒐𝒏𝒔 𝒅𝒆𝒖𝒙 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆𝒔 𝒂 > 𝟎 𝒆𝒕 𝜶 > 𝟎 . 𝑶𝒏 𝒅𝒊𝒓𝒂 𝒒𝒖𝒆 𝑿 𝒔𝒖𝒊𝒕 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒅𝒆 𝑷𝒂𝒓𝒆𝒕𝒐 𝒅𝒆
𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆𝒔 (𝒂 , 𝜶), 𝒔𝒊 𝑿 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒇 𝒅é𝒇𝒊𝒏𝒊𝒆 𝒑𝒂𝒓 ∶
𝜶 𝒂 𝜶+𝟏
𝒇𝑿 (𝒙) = {𝒂 (𝒙) , 𝒔𝒊 𝒙 ≥ 𝒂
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
𝑪𝒆𝒕𝒕𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒅é𝒄𝒓𝒊𝒕 𝒑𝒂𝒓 𝒆𝒙𝒆𝒎𝒑𝒍𝒆 𝒍𝒂 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆𝒔 𝒓𝒊𝒄𝒉𝒆𝒔𝒔𝒆𝒔.
1) 𝑴𝒐𝒏𝒕𝒓𝒆𝒓 𝒒𝒖𝒆 𝒇 𝒆𝒔𝒕 𝒃𝒊𝒆𝒏 𝒖𝒏𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒅𝒆 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é. 𝑨𝒍𝒍𝒖𝒓𝒆 𝒅𝒖 𝒈𝒓𝒂𝒑𝒉𝒆 𝒅𝒆 𝒇 ∶ 𝒐𝒏
𝒑𝒐𝒖𝒓𝒓𝒂 𝒑𝒓𝒆𝒏𝒅𝒓𝒆, 𝒂 = 𝟏 𝒆𝒕 𝜶 = 𝟑
2) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝑷(𝑿 > 𝒙) . 𝑨𝒍𝒍𝒖𝒓𝒆 𝒅𝒆 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒂𝒕𝒊𝒐𝒏 𝒑𝒐𝒖𝒓 𝒍𝒆𝒔 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆𝒔
𝒄𝒊‑𝒅𝒆𝒔𝒔𝒖𝒔.
3) 𝑺𝒐𝒊𝒕 𝒚 > 𝟎 𝒇𝒊𝒙é. 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝐥𝐢𝐦 𝑷(𝑿 > 𝒙 + 𝒚|𝑿 > 𝒙) . 𝑸𝒖’𝒆𝒏 𝒄𝒐𝒏𝒄𝒍𝒖𝒓𝒆 ?
𝒙→+∞
𝑪𝒆𝒕𝒕𝒆 𝒑𝒓𝒐𝒑𝒓𝒊é𝒕é 𝒆𝒔𝒕‑𝒆𝒍𝒍𝒆 𝒗𝒓𝒂𝒊𝒆 𝒑𝒐𝒖𝒓 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒆𝒙𝒑𝒐𝒏𝒆𝒏𝒕𝒊𝒆𝒍𝒍𝒆 ?
4) 𝑴𝒐𝒏𝒕𝒓𝒆𝒓 𝒒𝒖𝒆 𝑿 𝒂𝒅𝒎𝒆𝒕 𝒖𝒏𝒆 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 𝜶 > 𝟏 .
𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝑬(𝑿) 𝒅𝒂𝒏𝒔 𝒄𝒆 𝒄𝒂𝒔
5) 𝑴𝒐𝒏𝒕𝒓𝒆𝒓 𝒒𝒖𝒆 𝑿 𝒂𝒅𝒎𝒆𝒕 𝒖𝒏𝒆 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 𝜶 > 𝟐.
𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝑽(𝑿) 𝒅𝒂𝒏𝒔 𝒄𝒆 𝒄𝒂𝒔

Corrigé
+∞
∫ 𝒇𝑿 (𝒙)𝒅𝒙 = 𝟏 (𝟏)
1) 𝒇𝑿 (𝒙) 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒅. 𝒅. 𝒑 ⟺ { 𝒂
𝒇𝑿 (𝒙) ≥ 𝟎 , ∀𝒙 ∈ [𝒂, +∞[ (𝟐)
+∞ +∞ +∞
𝜶 𝒂 𝜶+𝟏 𝟏
(𝟏) ∶ ∫ 𝒇𝑿 (𝒙)𝒅𝒙 = ∫ ( ) 𝒅𝒙 = 𝜶(𝒂)𝜶 ∫ 𝒅𝒙
𝒂 𝒂 𝒂 𝒙 𝒂 𝒙(𝜶+𝟏)
+∞
𝟏
𝒐𝒓 𝜶 > 𝟎 ⟹ 𝜶 + 𝟏 > 𝟏 , 𝒑𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 ∫ 𝒅𝒙 , 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 , (𝑰𝒏𝒕é𝒈𝒓𝒂𝒍𝒆 𝒅𝒆 𝑹𝒊𝒆𝒎𝒂𝒏𝒏)
𝒂 𝒙(𝜶+𝟏)
+∞
+∞
𝜶
+∞
−(𝜶+𝟏)
𝒙−(𝜶+𝟏)+𝟏
𝜶 𝜶
𝟏 +∞
∫ 𝒇𝑿 (𝒙)𝒅𝒙 = 𝜶(𝒂) ∫ 𝒙 𝒅𝒙 = 𝜶(𝒂) [ ] = −(𝒂) [ 𝜶 ]
𝒂 𝒂 −(𝜶 + 𝟏) + 𝟏 𝒂 𝒙 𝒂

339 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
+∞
𝟏 𝟏
∫ 𝒇𝑿 (𝒙)𝒅𝒙 = −(𝒂)𝜶 [( 𝐥𝐢𝐦 𝜶 ) − 𝜶 ] = 𝟏 , 𝒂𝒊𝒏𝒔𝒊 (𝟏) 𝒆𝒔𝒕 𝒗é𝒆𝒊𝒇𝒊é𝒆
𝒂 ⏟𝒙→+∞ 𝒙 𝒂
𝟎

𝜶 𝒂 𝜶+𝟏
(𝟐) ∶ 𝒂 > 𝟎 𝒆𝒕 𝜶 > 𝟎, 𝒇𝑿 (𝒙) = ( ) ⇒ 𝒇𝑿 (𝒙) ≥ 𝟎 , ∀𝒙 ∈ [𝒂, +∞[ , 𝒂𝒊𝒏𝒔𝒊 (𝟐) 𝒆𝒔𝒕 𝒗é𝒆𝒊𝒇𝒊é𝒆
𝒂 𝒙
𝑫′ 𝒐ù 𝒇𝑿 (𝒙) 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒅. 𝒅. 𝒑

+∞
+∞ 𝟏 +∞ 𝟏 𝟏
2) 𝑭𝑿 (𝒙) = 𝟏 − 𝑷(𝑿 > 𝒙) = 𝟏 − ∫𝒙 𝒇𝑿 (𝒕)𝒅𝒕 = −(𝒂)𝜶 [𝒕𝜶 ] = 𝟏 + (𝒂)𝜶 [( 𝐥𝐢𝐦 𝒕𝜶 ) − 𝒙𝜶 ]
𝒙 ⏟𝒕→+∞
𝟎 𝒙
𝒂 𝜶
= 𝟏 − ( ) , 𝒔𝒊 𝒙 ≥ 𝒂
𝒙
𝟎 , 𝒔𝒊 𝒙 < 𝒂
′
𝑫 𝒐ù, 𝑭𝑿 (𝒙) = { 𝒂 𝜶
𝟏 − ( ) , 𝒔𝒊 𝒙 ≥ 𝒂
𝒙

𝑷[𝑿∈{](𝒙+𝒚),+∞[∩]𝒙,+∞[}]
3) 𝑷(𝑿 > 𝒙 + 𝒚|𝑿 > 𝒙) =
𝑷(𝑿>𝒙)

𝒐𝒓 𝒑𝒐𝒖𝒓, 𝒚 > 𝟎 , ](𝒙 + 𝒚), +∞[ ∩ ]𝒙, +∞[ = ](𝒙 + 𝒚), +∞[ , 𝒑𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 ∶
𝑷(𝑿 > 𝒙 + 𝒚) 𝟏 − 𝑭𝑿 (𝒙 + 𝒚) (𝒂⁄𝒙 + 𝒚)𝜶 𝒙 𝜶
𝑷(𝑿 > 𝒙 + 𝒚|𝑿 > 𝒙) = = = = ( )
𝑷(𝑿 > 𝒙) 𝟏 − 𝑭𝑿 (𝒙) (𝒂⁄𝒙)𝜶 𝒙+𝒚
𝒙 𝜶
𝑬𝒏 𝒆𝒇𝒇𝒆𝒕, 𝐥𝐢𝐦 𝑷(𝑿 > 𝒙 + 𝒚|𝑿 > 𝒙) = 𝐥𝐢𝐦 ( ) =𝟏
𝒙→+∞ 𝒙→+∞ 𝒙 + 𝒚

340 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
∙ 𝑪𝒆𝒍𝒂 𝒗𝒆𝒖𝒕 𝒅𝒊𝒓𝒆 𝒒𝒖𝒆 𝒑𝒍𝒖𝒔 𝑿 𝒑𝒓𝒆𝒏𝒅 𝒅𝒆 𝒈𝒓𝒂𝒏𝒅𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓𝒔, 𝒑𝒍𝒖𝒔 𝒆𝒍𝒍𝒆 𝒂 𝒅𝒆 𝒄𝒉𝒂𝒏𝒄𝒆𝒔 𝒅’𝒆𝒏 𝒑𝒓𝒆𝒏𝒅𝒓𝒆
𝒅𝒆 𝒑𝒍𝒖𝒔 𝒈𝒓𝒂𝒏𝒅𝒆𝒔. 𝑪𝒆𝒕𝒕𝒆 𝒍𝒐𝒊 𝒇𝒖𝒕 𝒊𝒏𝒕𝒓𝒐𝒅𝒖𝒊𝒕𝒆 𝒑𝒂𝒓 𝑷𝒂𝒓𝒆𝒕𝒐 𝒄𝒐𝒎𝒎𝒆 𝒎𝒐𝒅è𝒍𝒆 𝒅𝒆 𝒓𝒊𝒄𝒉𝒆𝒔𝒔𝒆, 𝒄𝒆𝒍𝒖𝒊‑𝒄𝒊

𝒂𝒚𝒂𝒏𝒕 𝒓𝒆𝒎𝒂𝒓𝒒𝒖é 𝒂𝒖 𝒅é𝒃𝒖𝒕 𝒅𝒖 𝟐𝟎è𝒎𝒆 𝒔𝒊è𝒄𝒍𝒆 𝒒𝒖𝒆 𝟐𝟎% 𝒅𝒆 𝒍𝒂 𝒑𝒐𝒑𝒖𝒍𝒂𝒕𝒊𝒐𝒏 𝒑𝒐𝒔𝒔é𝒅𝒂𝒊𝒕 𝟖𝟎% 𝒅𝒆𝒔
𝒓𝒊𝒄𝒉𝒆𝒔𝒔𝒆𝒔. 𝑫’𝒂𝒖𝒕𝒓𝒆𝒔 𝒑𝒉é𝒏𝒐𝒎è𝒏𝒆𝒔 𝒐𝒏𝒕 𝒄𝒆 𝒎ê𝒎𝒆 𝒕𝒚𝒑𝒆 𝒅𝒆 𝒑𝒓𝒐𝒑𝒓𝒊é𝒕é ∶ 𝒑𝒐𝒖𝒓 𝒖𝒏 𝒔𝒆𝒓𝒗𝒊𝒄𝒆, 𝟐𝟎%
𝒅𝒆𝒔 𝒄𝒍𝒊𝒆𝒏𝒕𝒔 𝒔𝒐𝒏𝒕 𝒓𝒆𝒔𝒑𝒐𝒏𝒔𝒂𝒃𝒍𝒆𝒔 𝒅𝒆 𝟖𝟎% 𝒅𝒆𝒔 𝒓é𝒄𝒍𝒂𝒎𝒂𝒕𝒊𝒐𝒏𝒔 …
∙ 𝑪𝒆𝒍𝒂 𝒏’𝒆𝒔𝒕 𝒑𝒂𝒔 𝒗𝒓𝒂𝒊 𝒑𝒐𝒖𝒓 𝒍𝒂 𝒍𝒐𝒊 𝒆𝒙𝒑𝒐𝒏𝒆𝒏𝒕𝒊𝒆𝒍𝒍𝒆 𝒒𝒖𝒊 𝒏’𝒂 𝒑𝒂𝒔 𝒅𝒆 𝒎é𝒎𝒐𝒊𝒓𝒆 ∶
𝒂 𝜶
∀𝒙, 𝒚 > 𝟎 ∶ 𝑷(𝑿 > 𝒙 + 𝒚|𝑿 > 𝒙) = 𝑷(𝑿 > 𝒚) = 𝟏 − 𝑭𝑿 (𝒚) = ( )
𝒚
𝒂 𝜶
𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆, 𝐥𝐢𝐦 𝑷(𝑿 > 𝒙 + 𝒚|𝑿 > 𝒙) = ( )
𝒙→+∞ 𝒚
+∞ +∞ 𝟏
4) 𝑬(𝑿) = ∫𝒂 𝒙𝒇𝑿 (𝒙)𝒅𝒙 = 𝜶(𝒂)𝜶 ∫𝒂 𝒅𝒙 , 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒔𝒊 𝜶 > 𝟏, (𝑰𝒏𝒕é𝒈𝒓𝒂𝒍𝒆 𝒅𝒆 𝑹𝒊𝒆𝒎𝒂𝒏𝒏)
𝒙𝜶

𝑬(𝑿) 𝒆𝒙𝒊𝒔𝒕𝒆, 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 𝜶 > 𝟏

+∞
𝜶
+∞
−𝜶 𝜶
𝒙−𝜶+𝟏 𝜶(𝒂)𝜶 𝟏 +∞ 𝜶(𝒂)𝜶 𝟏 𝟏
𝑬(𝑿) = 𝜶(𝒂) ∫ 𝒙 𝒅𝒙 = 𝜶(𝒂) [ ] =− [ 𝜶−𝟏 ] = [ 𝜶−𝟏 − ( 𝐥𝐢𝐦 𝜶−𝟏 )]
𝒂 −𝜶 + 𝟏 𝒂 𝜶−𝟏 𝒙 𝒂 𝜶−𝟏 𝒂 ⏟𝒙→+∞ 𝒙
𝟎

𝜶(𝒂)𝜶 𝜶𝒂
= 𝜶−𝟏
, 𝒅′ 𝒐ù ∶ 𝒑𝒐𝒖𝒓, 𝜶 > 𝟏 ∶ 𝑬(𝑿) =
(𝜶 − 𝟏)𝒂 𝜶−𝟏
+∞ +∞ 𝟏
5) 𝑬(𝑿𝟐 ) = ∫𝒂 𝒙𝟐 𝒇𝑿 (𝒙)𝒅𝒙 = 𝜶(𝒂)𝜶 ∫𝒂 𝒅𝒙 , 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒔𝒊 𝜶 − 𝟏 > 𝟏 𝒐𝒖 𝒆𝒏𝒄𝒓𝒆 , 𝒔𝒊 𝜶 > 𝟐
𝒙𝜶−𝟏
+∞
𝟐) 𝜶
+∞
𝟏−𝜶
𝒙𝟏−𝜶+𝟏
𝜶
𝜶(𝒂)𝜶 𝟏 +∞
𝑬(𝑿 = 𝜶(𝒂) ∫ 𝒙 𝒅𝒙 = 𝜶(𝒂) [ ] =− [ ]
𝒂 𝟏−𝜶+𝟏 𝒂 𝜶 − 𝟐 𝒙𝜶−𝟐 𝒂

𝜶(𝒂)𝜶 𝟏 𝟏 𝜶(𝒂)𝜶 𝜶𝒂𝟐

= [ − ( 𝐥𝐢𝐦 )] = =
𝜶 − 𝟐 𝒂𝜶−𝟐 ⏟𝒙→+∞ 𝒙𝜶−𝟐 (𝜶 − 𝟐)𝒂𝜶−𝟐 𝜶 − 𝟐
𝟎

𝜶𝒂𝟐 𝜶𝟐 𝒂𝟐 𝟏 𝜶
𝑬𝒏 𝒆𝒇𝒇𝒆𝒕 ∶ 𝑽(𝑿) = 𝑬(𝑿𝟐 ) − [𝑬(𝑿)]𝟐 = − = 𝜶𝒂𝟐
[ − ]
𝜶 − 𝟐 (𝜶 − 𝟏)𝟐 𝜶 − 𝟐 (𝜶 − 𝟏)𝟐
(𝜶 − 𝟏)𝟐 − 𝜶(𝜶 − 𝟐) 𝜶𝟐 − 𝟐𝜶 + 𝟏 − 𝜶𝟐 + 𝟐𝜶
= 𝜶𝒂𝟐 [ ] = 𝜶𝒂𝟐
[ ]
(𝜶 − 𝟐)(𝜶 − 𝟏)𝟐 (𝜶 − 𝟐)(𝜶 − 𝟏)𝟐

′
𝜶𝒂𝟐
𝑫 𝒐ù ∶ 𝒑𝒐𝒖𝒓, 𝜶 > 𝟐 ∶ 𝑽(𝑿) =
(𝜶 − 𝟐)(𝜶 − 𝟏)𝟐

341 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
Exercice 19 : (Maximum de variables aléatoires continues et indépendantes)
ÉNONCÉ
𝑼𝒏 𝒂𝒏𝒂𝒍𝒚𝒔𝒕𝒆 𝒇𝒊𝒏𝒂𝒏𝒄𝒊𝒆𝒓 𝒅𝒊𝒔𝒑𝒐𝒔𝒆 𝒅𝒆 𝒅𝒐𝒏𝒏é𝒆𝒔 𝒉𝒊𝒔𝒕𝒐𝒓𝒊𝒒𝒖𝒆𝒔 𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 𝒅𝒆 𝒓𝒆𝒏𝒅𝒆𝒎𝒆𝒏𝒕 𝒅’𝒖𝒏𝒆
𝒄𝒆𝒓𝒕𝒂𝒊𝒏𝒆 𝒂𝒄𝒕𝒊𝒐𝒏. 𝑶𝒏 𝒔𝒖𝒑𝒑𝒐𝒔𝒆 𝒑𝒂𝒓 𝒔𝒊𝒎𝒑𝒍𝒊𝒄𝒊𝒕é 𝒒𝒖𝒆 𝒄𝒆𝒔 𝒐𝒃𝒔𝒆𝒓𝒗𝒂𝒕𝒊𝒐𝒏𝒔 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒂𝒗𝒆𝒄
𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝑭𝑿 (𝒙) 𝒆𝒕 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒇𝑿 (𝒙). 𝑺𝒐𝒊𝒕 𝑻 = 𝐦𝐚𝐱(𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 ) 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆
𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒒𝒖𝒊 𝒓𝒆𝒑𝒓é𝒔𝒆𝒏𝒕𝒆 𝒍𝒆 𝒓𝒆𝒏𝒅𝒆𝒎𝒆𝒏𝒕 𝒎𝒂𝒙𝒊𝒎𝒂𝒍.
1) 𝑴𝒐𝒏𝒕𝒓𝒆𝒓 𝒒𝒖𝒆, 𝑭𝑻 (𝒕) = [𝑭𝑿 (𝒕)]𝒏 , 𝒐ù 𝑭𝑻 (𝒕) 𝒆𝒔𝒕 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝑻 .
𝟐
𝒙 , 𝒔𝒊 𝟎 < 𝒙 < 𝜽
2) 𝑺𝒐𝒊𝒕 : 𝒇𝑿 (𝒙) = { 𝜽𝟐
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
𝑫𝒂𝒏𝒔 𝒄𝒆 𝒄𝒂𝒔, 𝒄𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝑬(𝑻) .
3) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒂 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒒𝒖𝒆 𝒍𝒆 𝒓𝒆𝒏𝒅𝒆𝒎𝒆𝒏𝒕 𝒎𝒂𝒙𝒊𝒎𝒂𝒍 𝒔𝒐𝒊𝒕 𝒔𝒖𝒑é𝒓𝒊𝒆𝒖𝒓 à 𝒖𝒏 𝒄𝒆𝒓𝒕𝒂𝒊𝒏 𝒔𝒆𝒖𝒊𝒍 𝒂.

Corrigé
1) 𝑬𝒕𝒂𝒏𝒕 𝒅𝒐𝒏𝒏é (𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 ) 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒆𝒕 𝒅𝒆 𝒎ê𝒎𝒆 𝒍𝒐𝒊 ; 𝒐𝒏 𝒂 ∶
𝒏

𝑭𝑻 (𝒕) = 𝑷(𝑻 ≤ 𝒕) = 𝑷(𝐦𝐚𝐱(𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 ) ≤ 𝒕) = 𝑷(𝑿𝟏 ≤ 𝒕, 𝑿𝟐 ≤ 𝒕 , … , 𝑿𝒏 ≤ 𝒕) = ∏ 𝑷(𝑿𝒊 ≤ 𝒕)

𝒊=𝟏
𝒏 𝒏

= ∏ 𝑭𝑿𝒊 (𝒕) = ∏ 𝑭𝑿 (𝒙) = [𝑭𝑿 (𝒕)]𝒏

𝒊=𝟏 𝒊=𝟏

𝑫′ 𝒐ù 𝑭𝑻 (𝒕) = [𝑭𝑿 (𝒕)]𝒏

2) 𝑳𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝑻 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏 ∶

𝒏
𝒅[𝑭𝑻 (𝒕)] 𝒅[(𝑭𝑿 (𝒕)) ] 𝒏−𝟏 𝒏−𝟏
∀𝒕 ∈ ]𝟎, 𝜽[ ; 𝒇𝑻 (𝒕) = = = 𝒏𝑭′𝑿 (𝒕)(𝑭𝑿 (𝒕)) = 𝒏𝒇𝑿 (𝒕)(𝑭𝑿 (𝒕))
𝒅𝒕 𝒅𝒕
𝟎 , 𝒔𝒊 𝒙 ≤ 𝟎
𝟐 𝒕
𝟐
𝑶𝒓 𝒐𝒏 𝒂 ∶ 𝒇𝑿 (𝒕) = { 𝜽𝟐
𝒕 , 𝒔𝒊 𝟎 < 𝒕 < 𝜽 ⇒ 𝑭𝑿 (𝒕) = ∫ 𝟐 𝒙𝒅𝒙 , 𝒔𝒊 𝟎 < 𝒕 < 𝜽
𝟎 , 𝒔𝒊 𝒏𝒐𝒏 𝟎 𝜽
{ 𝟏 , 𝒔𝒊 𝒕 ≥ 𝜽
𝟎 , 𝒔𝒊 𝒙 ≤ 𝟎 𝟎 , 𝒔𝒊 𝒙 ≤ 𝟎
𝒕
𝟐 𝒙𝟐 𝒕 𝟐
𝑭𝑿 ( 𝒕) = [ ] , 𝒔𝒊 𝟎 < 𝒕 < 𝜽 ⇒ 𝑭 𝑿 ( 𝒕 ) = { ( ) , 𝒔𝒊 𝟎 < 𝒕 < 𝜽
𝜽𝟐 𝟐 𝟎 𝜽
{ 𝟏 , 𝒔𝒊 𝒕 ≥ 𝜽 𝟏 , 𝒔𝒊 𝒕 ≥ 𝜽
𝒏−𝟏
𝒏−𝟏 𝟐𝒕 𝒕 𝟐 𝟐𝒏𝒕 𝒕𝟐(𝒏−𝟏)
𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆, ∀𝒕 ∈ ]𝟎, 𝜽[ ; 𝒇𝑻 (𝒕) = 𝒏𝒇𝑿 (𝒕)(𝑭𝑿 (𝒕)) = 𝒏 ( 𝟐 ) (( ) ) = ( 𝟐 ) ( 𝟐(𝒏−𝟏) )
𝜽 𝜽 𝜽 𝜽

𝟐𝒏𝒕𝟐𝒏−𝟏
𝑫′ 𝒐ù, 𝒇𝑻 (𝒕) = { 𝜽𝟐𝒏 , 𝒔𝒊 𝟎 < 𝒕 < 𝜽
𝟎 , 𝒔𝒊 𝒏𝒐𝒏

342 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝜽 𝜽 𝜽
𝟐𝒏𝒕𝟐𝒏 𝟐𝒏 𝜽 𝟐𝒏 𝟐𝒏 𝒕𝟐𝒏+𝟏
𝑬𝒏𝒇𝒊𝒏, 𝑬(𝑻) = ∫ 𝒕𝒇𝑻 (𝒕)𝒅𝒕 =∫ 𝒅𝒕 = 𝟐𝒏 ∫ 𝒕 𝒅𝒕 = 𝟐𝒏 [ ]
𝟎 𝟎 𝜽𝟐𝒏 𝜽 𝟎 𝜽 𝟐𝒏 + 𝟏 𝟎

𝟐𝒏
= (𝜽𝟐𝒏+𝟏 − 𝟎)
(𝟐𝒏 + 𝟏)𝜽𝟐𝒏

𝟐𝒏
𝑫′ 𝒐ù, 𝑬(𝑻) = ( )𝜽
𝟐𝒏 + 𝟏
3) 𝑶𝒏 𝒔𝒆 𝒑𝒓𝒐𝒑𝒐𝒔𝒆 𝒅𝒆 𝒄𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒂 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é :
𝑷(𝑻 > 𝒂) = 𝟏 − 𝑷(𝑻 ≤ 𝒂) = 𝟏 − 𝑭𝑻 (𝒂) = 𝟏 − [𝑭𝑿 (𝒂)]𝒏

𝟏 , 𝒔𝒊 𝒙 ≤ 𝟎
𝒕 𝟐𝒏
𝑫′ 𝒐ù, 𝑷(𝑻 > 𝒂) = {𝟏 − ( ) , 𝒔𝒊 𝟎 < 𝒕 < 𝜽
𝜽
𝟎 , 𝒔𝒊 𝒕 ≥ 𝜽

Exercice 20 : (Inégalité de Tchebychev)

ÉNONCÉ
𝑳𝒂 𝒇𝒍𝒖𝒄𝒕𝒖𝒂𝒕𝒊𝒐𝒏 𝒋𝒐𝒖𝒓𝒏𝒂𝒍𝒊è𝒓𝒆 𝒅𝒖 𝒑𝒓𝒊𝒙 𝒅𝒆 𝒍’𝒂𝒄𝒕𝒊𝒐𝒏 𝒅’𝒖𝒏𝒆 𝒔𝒐𝒄𝒊é𝒕é 𝒅𝒐𝒏𝒏é𝒆, 𝒄𝒐𝒕é𝒆 𝒆𝒏 𝒃𝒐𝒖𝒓𝒔𝒆, 𝒆𝒔𝒕 𝒖𝒏𝒆
𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆 𝒅’𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝟎 𝒆𝒕 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝝈𝟐 . 𝑪𝒆𝒍𝒂 𝒗𝒆𝒖𝒕 𝒅𝒊𝒓𝒆 𝒒𝒖𝒆, 𝒔𝒊 𝒀𝒏 𝒓𝒆𝒑𝒓é𝒔𝒆𝒏𝒕𝒆 𝒍𝒆
𝒑𝒓𝒊𝒙 𝒅𝒆 𝒍’𝒂𝒄𝒕𝒊𝒐𝒏 𝒅𝒖 𝒏è𝒎𝒆 𝒋𝒐𝒖𝒓 , 𝒂𝒍𝒐𝒓𝒔 𝒀𝒏 = 𝒀𝒏−𝟏 + 𝑼𝒏 ; 𝒏 > 𝟏 𝒐ù 𝑼𝟏 , 𝑼𝟐 , … , 𝑼𝒏 𝒔𝒐𝒏𝒕 𝒅𝒆𝒔
𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒊𝒅𝒆𝒏𝒕𝒊𝒒𝒖𝒆𝒎𝒆𝒏𝒕 𝒅𝒊𝒔𝒕𝒓𝒊𝒃𝒖é𝒆𝒔 𝒅’𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝟎 𝒆𝒕 𝒅𝒆
𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝝈𝟐 . 𝑺𝒖𝒑𝒑𝒐𝒔𝒐𝒏𝒔 𝒒𝒖𝒆 𝒍𝒆 𝒑𝒓𝒊𝒙 𝒅𝒆 𝒍’𝒂𝒄𝒕𝒊𝒐𝒏 𝒔𝒐𝒊𝒕 𝒂𝒖𝒋𝒐𝒖𝒓𝒅’𝒉𝒖𝒊 𝒅𝒆 𝟏𝟎𝟎, 𝒄‑à‑𝒅. 𝒀𝟏 = 𝒚𝟏 = 𝟏𝟎𝟎
𝒆𝒕 𝒒𝒖𝒆 𝝈𝟐 = 𝟏 .
𝑫𝒐𝒏𝒏𝒆𝒓 𝒖𝒏𝒆 𝒃𝒐𝒓𝒏𝒆 𝒊𝒏𝒇𝒆𝒓𝒊𝒆𝒖𝒓𝒆 𝒑𝒐𝒖𝒓 𝒍𝒂 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒒𝒖𝒆 𝒍𝒆 𝒑𝒓𝒊𝒙 𝒅𝒆 𝒍’𝒂𝒄𝒕𝒊𝒐𝒏 𝒔𝒆𝒓𝒂 𝒄𝒐𝒎𝒑𝒓𝒊𝒔
𝒆𝒏𝒕𝒓𝒆 𝟗𝟓 𝒆𝒕 𝟏𝟎𝟓 𝒅𝒂𝒏𝒔 𝟏𝟎 𝒋𝒐𝒖𝒓𝒔, 𝒆𝒏 𝒖𝒕𝒊𝒍𝒊𝒔𝒂𝒏𝒕 𝒍’𝒊𝒏é𝒈𝒂𝒍𝒊𝒕é 𝒅𝒆 𝑻𝒄𝒉𝒆𝒃𝒚𝒄𝒉𝒆𝒗.

Corrigé
∙ 𝑶𝒏 𝒂 ∶ 𝒀𝟐 = 𝒀𝟏 + 𝑼𝟐 , 𝒀𝟑 = 𝒀𝟐 + 𝑼𝟑 = 𝒀𝟏 + 𝑼𝟐 + 𝑼𝟑 , 𝒀𝟒 = 𝒀𝟑 + 𝑼𝟒 = 𝒀𝟏 + 𝑼𝟐 + 𝑼𝟑 + 𝑼𝟒 …
𝟏𝟏

𝑳𝒆 𝒑𝒓𝒊𝒙 𝒅𝒆 𝒍’𝒂𝒄𝒕𝒊𝒐𝒏 𝒅𝒂𝒏𝒔 𝟏𝟎 𝒋𝒐𝒖𝒓𝒔 𝒔’é𝒄𝒓𝒊𝒕 ∶ 𝒀𝟏𝟏 = 𝒀𝟏 + ∑ 𝑼𝒊

𝒊=𝟐
𝟏𝟏 𝟏𝟏 𝟏𝟏

∙ 𝑺𝒐𝒏 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒆𝒔𝒕 𝒂𝒍𝒐𝒓𝒔 ∶ 𝑬(𝒀𝟏𝟏 ) = 𝑬 (𝒀𝟏 + ∑ 𝑼𝒊 ) = 𝑬 (𝟏𝟎𝟎 + ∑ 𝑼𝒊 ) = 𝟏𝟎𝟎 + 𝑬 (∑ 𝑼𝒊 )

𝒊=𝟐 𝒊=𝟐 𝒊=𝟐
𝟏𝟏

𝑬(𝒀𝟏𝟏 ) = 𝟏𝟎𝟎 + ∑ 𝑬(𝑼

⏟ 𝒊 ) = 𝟏𝟎𝟎
𝒊=𝟐 𝟎

343 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝟏𝟏 𝟏𝟏 𝟏𝟏

∙ 𝑬𝒕 𝒔𝒂 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 ∶ 𝑽(𝒀𝟏𝟏 ) = 𝑽 (𝒀𝟏 + ∑ 𝑼𝒊 ) = 𝑽 (𝟏𝟎𝟎 + ∑ 𝑼𝒊 ) = 𝑽 (∑ 𝑼𝒊 )

𝒊=𝟐 𝒊=𝟐 𝒊=𝟐
𝟏𝟏 𝟏𝟏

⏟ 𝒊 ) = 𝟏𝟎𝝈𝟐 = 𝟏𝟎
𝑶𝒓 𝒍𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝑼𝟏 , 𝑼𝟐 , … , 𝑼𝒏 𝒔𝒐𝒏𝒕 𝒊. 𝒊. 𝒅 , 𝒅𝒐𝒏𝒄 𝑽 (∑ 𝑼𝒊 ) = ∑ 𝑽(𝑼
𝒊=𝟐 𝒊=𝟐 𝝈𝟐

𝑨𝒊𝒏𝒔𝒊, 𝑽(𝒀𝟏𝟏 ) = 𝟏𝟎
∙ 𝑷(𝟗𝟓 < 𝒀𝟏𝟏 < 𝟏𝟎𝟓) = 𝑷(−𝟓 < 𝒀𝟏𝟏 − 𝟏𝟎𝟎 < 𝟓) = 𝑷(|𝒀𝟏𝟏 − 𝑬(𝒀𝟏𝟏 )| < 𝟓)
= 𝟏 − 𝑷(|𝒀𝟏𝟏 − 𝑬(𝒀𝟏𝟏 )| ≥ 𝟓)
𝑽𝒂𝒓(𝒀𝟏𝟏 )
𝑶𝒓 𝒑𝒂𝒓 𝒍’𝒊𝒏é𝒈𝒂𝒍𝒊𝒕é 𝒅𝒆 𝑻𝒄𝒉𝒆𝒃𝒚𝒄𝒉𝒆𝒗 ∶ 𝑷(|𝒀𝟏𝟏 − 𝑬(𝒀𝟏𝟏 )| ≥ 𝟓) ≤
𝟓𝟐
𝟏𝟎
⟺ 𝑷(|𝒀𝟏𝟏 − 𝑬(𝒀𝟏𝟏 )| ≥ 𝟓) ≤ ⟺ −𝑷(|𝒀𝟏𝟏 − 𝑬(𝒀𝟏𝟏 )| ≥ 𝟓) ≥ −𝟎, 𝟒 ⟺ 𝟏 − 𝑷(|𝒀𝟏𝟏 − 𝑬(𝒀𝟏𝟏 )| ≥ 𝟓) ≥ 𝟎, 𝟔
𝟐𝟓

𝒆𝒕 𝟎, 𝟔 ≤ 𝑷(𝟗𝟓 < 𝒀𝟏𝟏 < 𝟏𝟎𝟓)

𝒄‑à‑𝒅. 𝒒𝒖’𝒊𝒍 𝒚 𝒂, 𝒂𝒖 𝒎𝒐𝒊𝒏𝒔 𝟔𝟎 % 𝒅𝒆 𝒄𝒉𝒂𝒏𝒄𝒆 𝒒𝒖𝒆 𝒍𝒆 𝒑𝒓𝒊𝒙 𝒅𝒆 𝒍’𝒂𝒄𝒕𝒊𝒐𝒏 𝒔𝒆 𝒕𝒓𝒐𝒖𝒗𝒆 𝒆𝒏𝒕𝒓𝒆 𝟗𝟓 𝒆𝒕 𝟏𝟎𝟓
𝒅𝒂𝒏𝒔 𝟏𝟎 𝒋𝒐𝒖𝒓𝒔.

Exercice 21 : (Transformations de variables-Loi lo‑normale-Modes de convergence)

ÉNONCÉ
𝑪𝒆 𝒑𝒓𝒐𝒃𝒍è𝒎𝒆 𝒆𝒔𝒕 𝒃𝒂𝒔é 𝒔𝒖𝒓 𝒍’𝒂𝒓𝒕𝒊𝒄𝒍𝒆 « 𝑻𝒉𝒆 𝑳𝒐𝒏𝒈‑𝑻𝒆𝒓𝒎 𝑬𝒙𝒑𝒆𝒄𝒕𝒆𝒅 𝑹𝒂𝒕𝒆 𝒐𝒇 𝑹𝒆𝒕𝒖𝒓𝒏 ∶ 𝑺𝒆𝒕𝒕𝒊𝒏𝒈 𝒊𝒕 𝑹𝒊𝒈𝒉𝒕 », 𝒑𝒖𝒃𝒍𝒊é 𝒑𝒂𝒓

𝑶. 𝒅𝒆 𝒍𝒂 𝑮𝒓𝒂𝒏𝒅𝒗𝒊𝒍𝒍𝒆 𝒅𝒂𝒏𝒔 𝒍𝒆 𝑭𝒊𝒏𝒂𝒏𝒄𝒊𝒂𝒍 𝑨𝒏𝒂𝒍𝒚𝒔𝒕𝒔 𝑱𝒐𝒖𝒓𝒏𝒂𝒍 (𝟏𝟗𝟗𝟖, 𝒑𝒂𝒈𝒆𝒔 𝟕𝟓‑𝟖𝟎).

𝑺𝒐𝒊𝒕 𝑺𝒕 𝒍𝒂 𝒗𝒂𝒍𝒆𝒖𝒓 𝒅’𝒖𝒏 𝒂𝒄𝒕𝒊𝒇 à 𝒍𝒂 𝒇𝒊𝒏 𝒅𝒆 𝒍’𝒂𝒏𝒏é𝒆 𝒕 𝒆𝒕 𝑹𝟎,𝒏 𝒍𝒆 𝒕𝒂𝒖𝒙 𝒅𝒆 𝒓𝒆𝒏𝒅𝒆𝒎𝒆𝒏𝒕 𝒔𝒖𝒓 𝒖𝒏 𝒉𝒐𝒓𝒊𝒛𝒐𝒏
𝒏
𝒅𝒆 𝒏 𝒂𝒏𝒏é𝒆𝒔, 𝒄‑à‑𝒅. 𝒒𝒖𝒆 𝑹𝟎,𝒏 𝒆𝒔𝒕 𝒍𝒂 𝒔𝒐𝒍𝒖𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍′ é𝒒𝒖𝒂𝒕𝒊𝒖𝒐𝒏 ∶ 𝑺𝒏 = 𝑺𝟎 (𝟏 + 𝑹𝟎,𝒏 ) .
𝑺𝒕
𝑺𝒐𝒖𝒔 𝒍’𝒉𝒚𝒑𝒐𝒕𝒉è𝒔𝒆 𝒒𝒖𝒆 ( ) 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒊. 𝒊. 𝒅 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒍𝒐𝒈‑𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝓛𝓝(𝝁, 𝝈𝟐 )
𝑺𝒕−𝟏 𝒕∈ℕ∗

1) 𝑽é𝒓𝒊𝒇𝒊𝒆𝒓 𝒒𝒖𝒆 𝒔𝒊 𝑿 ↝ 𝓛𝓝(𝝁, 𝝈𝟐 ) , 𝒂𝒍𝒐𝒓𝒔 𝒀 = 𝐥𝐧 𝑿 ↝ 𝓝(𝝁, 𝝈𝟐 )

2)
a) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍’𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒎𝒂𝒕𝒉é𝒎𝒂𝒕𝒊𝒒𝒖𝒆 𝒅𝒆 𝑿
b) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒅𝒆 𝑿
3) 𝑫é𝒅𝒖𝒊𝒓𝒆 𝒂𝒍𝒐𝒓𝒔 𝒍’𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒆𝒕 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒅𝒆 𝑹𝟎,𝒏
4) 𝑸𝒖𝒆 𝒔𝒆 𝒑𝒂𝒔𝒔𝒆‑𝒕‑𝒊𝒍 𝒒𝒖𝒂𝒏𝒅 𝒏 → +∞
𝟏 [𝐥𝐧(𝒆𝒚 ) − 𝝁]𝟐
𝐞𝐱𝐩 (− ) , 𝐬𝐢 𝒙 > 𝟎
𝑰𝒏𝒅𝒊𝒄𝒂𝒕𝒊𝒐𝒏 ∶ 𝑿 ↝ 𝓛𝓝(𝝁, 𝝈𝟐 ) ⇒ 𝒇𝑿 (𝒙) = {𝝈𝒙√𝟐𝝅 𝟐𝝈𝟐
𝟎 , 𝒔𝒊 𝒏𝒐𝒏

344 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
Corrigé
𝑿(𝛀) = ]𝟎, +∞[
1) { ⇒ 𝒀(𝛀) = ℝ
𝒀 = 𝐥𝐧 𝑿
𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒁 :
𝑭𝒀 (𝒚) = 𝑷(𝒀 ≤ 𝒚) = 𝑷(𝐥𝐧 𝑿 ≤ 𝒚) = 𝑷(𝑿 ≤ 𝒆𝒚 ) = 𝑭𝑿 (𝒆𝒚 )
𝑳𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏 ∶

𝒅[𝑭𝑿 (𝒆𝒚 )] 𝟏 (𝐥𝐧(𝒆𝒚 ) − 𝝁)𝟐 𝟏 (𝒚−𝝁)𝟐

𝒚 )′ 𝒚) 𝒚 −
𝒇𝒀 (𝒚) = = (𝒆 𝒇𝑿 (𝒆 =𝒆 [ 𝐞𝐱𝐩 (− )] = 𝒆 𝟐𝝈𝟐
𝒅𝒚 𝝈𝒆𝒚 √𝟐𝝅 𝟐𝝈𝟐 𝝈√𝟐𝝅

𝑫′ 𝒐ù, 𝑿 ↝ 𝓛𝓝(𝝁, 𝝈𝟐 ) ⟹ 𝒀 = 𝐥𝐧 𝑿 ↝ 𝓝(𝝁, 𝝈𝟐 )

2)
a) 𝒀 = 𝐥𝐧 𝑿 ⟺ 𝑿 = 𝒆𝒀
𝑬(𝑿) = 𝑬(𝒆𝒀 ) = 𝑴𝒀 (𝟏) , 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒀 𝒆𝒏 𝟏
𝒀−𝝁
𝑫′ 𝒂𝒖𝒕𝒓𝒆 𝒑𝒂𝒓𝒕 𝒀 ↝ 𝓝(𝝁, 𝝈𝟐 ) ⟹ 𝒁 = ↝ 𝓝(𝟎, 𝟏)
𝝈
𝒕𝟐 𝝈𝟐 𝝈𝟐
(𝝁+ )
{𝑴𝒁 (𝒕) = 𝐞𝟐
⟹ 𝑴𝒀 (𝒕) = 𝑴𝝈𝒁+𝝁 (𝒕) = 𝒆𝝁𝒕 𝑴𝒁 (𝝈𝒕) ⟹ 𝑴𝒀 (𝟏) = 𝒆𝝁 𝑴𝒁 (𝝈) = 𝒆𝝁 𝐞 𝟐 =𝒆 𝟐

𝒀 = 𝝈𝒁 + 𝝁

𝝈𝟐
(𝝁+ )
𝟐
𝑫′ 𝒐ù, 𝑿 ↝ 𝓛𝓝(𝝁, 𝝈𝟐 ) ⟹ 𝑬(𝑿) = 𝒆

b)
(𝟐𝝈)𝟐 𝟐 +𝝁)
∙ 𝑬(𝑿𝟐 ) = 𝑬[(𝒆𝒀 )𝟐 ] = 𝑬(𝒆𝟐𝒀 ) = 𝑴𝒀 (𝟐) = 𝒆𝟐𝝁 𝑴𝒁 (𝟐𝝈) = 𝒆𝟐𝝁 𝒆 𝟐 = 𝒆𝟐(𝝈
𝟐
𝝈𝟐
(𝝁+ )
𝟐(𝝈𝟐 +𝝁) 𝟐 𝟐 𝟐) 𝟐 +𝝁) 𝟐
∙ 𝑽(𝑿) = 𝑬(𝑿𝟐 ) − [𝑬(𝑿)]𝟐 = 𝒆 − [𝒆 ] = 𝒆(𝟐𝝁+𝟐𝝈 ) − 𝒆(𝟐𝝁+𝝈 = 𝒆𝟐(𝝈 (𝟏 − 𝒆−𝝈 )

𝟐 +𝝁) 𝟐
𝑫′ 𝒐ù, 𝑿 ↝ 𝓛𝓝(𝝁, 𝝈𝟐 ) ⟹ 𝑽(𝑿) = 𝒆𝟐(𝝈 (𝟏 − 𝒆−𝝈 )

3)
𝟏 𝟏
𝒏 𝑺𝒏 𝒏 𝑺𝒏 𝒏 𝑺𝒏 𝒏
𝑺𝒏 = 𝑺𝟎 (𝟏 + 𝑹𝟎,𝒏 ) ⟺ (𝟏 + 𝑹𝟎,𝒏 ) = ⟺ 𝟏 + 𝑹𝟎,𝒏 = ( ) ⟺ 𝑹𝟎,𝒏 = ( ) − 𝟏
𝑺𝟎 𝑺𝟎 𝑺𝟎
𝒏
𝑺𝒕 𝑺𝒏
𝑬𝒏 𝒖𝒕𝒊𝒍𝒊𝒔𝒂𝒏𝒕 𝒍𝒂 𝒑𝒓𝒐𝒑𝒓𝒊é𝒕é 𝒅𝒆𝒔 𝒑𝒓𝒐𝒅𝒖𝒊𝒕𝒔 𝒕é𝒍𝒆𝒔𝒄𝒐𝒑𝒊𝒒𝒖𝒆𝒔, 𝒐𝒏 𝒐𝒃𝒕𝒊𝒆𝒏𝒕 ∶ ∏ =
𝑺𝒕−𝟏 𝑺𝟎
𝒕=𝟏
𝟏
𝒏 𝒏 𝟏 𝒏 𝟏
𝒏
𝑺𝒕 𝑺𝒕 𝒏 𝑺𝒕 𝒏
𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆, 𝑹𝟎,𝒏 = (∏ ) − 𝟏 = ∏( ) − 𝟏 = 𝐞𝐱𝐩 (𝐥𝐧 (∏ ( ) )) − 𝟏
𝑺𝒕−𝟏 𝑺𝒕−𝟏 𝑺𝒕−𝟏
𝒕=𝟏 𝒕=𝟏 𝒕=𝟏

345 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒏 𝟏 𝒏 𝒏
𝑺𝒕 𝒏 𝟏 𝑺𝒕 𝟏
𝑹𝟎,𝒏 = 𝐞𝐱𝐩 (∑ 𝐥𝐧 ( ) ) − 𝟏 = 𝐞𝐱𝐩 ( ∑ 𝐥𝐧 ( )) − 𝟏 = 𝐞𝐱𝐩 ( ∑ 𝑸𝒕 ) − 𝟏 = 𝒆𝑸̅ − 𝟏
𝑺𝒕−𝟏 𝒏 𝑺𝒕−𝟏 𝒏
𝒕=𝟏 𝒕=𝟏 𝒕=𝟏

𝑺𝒕
𝑨𝒗𝒆𝒄 , 𝑸𝒕 = 𝐥𝐧 ( )
𝑺𝒕−𝟏
𝑺𝒕 𝑺𝒕
𝑪𝒐𝒎𝒎𝒆 𝒐𝒏𝒂 ∶ ( ) ↝ 𝓛𝓝(𝝁, 𝝈𝟐 ) ⟹ 𝑸𝒕 = 𝐥𝐧 ( ) ↝ 𝓝(𝝁, 𝝈𝟐 )
𝑺𝒕−𝟏 𝑺𝒕−𝟏
𝑺𝒕 𝑺𝒕
𝑺𝒐𝒊𝒕 𝒇(𝒖) = 𝐥𝐧(𝒖) , 𝒅𝒐𝒏𝒄 𝑸𝒕 = 𝒇 ( ) , 𝒇 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒔𝒖𝒓 ( ) (𝛀) = ]𝟎, +∞[ (𝟏)
𝑺𝒕−𝟏 𝑺𝒕−𝟏
𝑺𝒕
( ) 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒊. 𝒊. 𝒅 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒍𝒐𝒈‑𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝓛𝓝(𝝁, 𝝈𝟐 ) (𝟐)
𝑺𝒕−𝟏 𝒕∈ℕ∗

𝑺𝒕
𝑬 (𝒇 ( )) = 𝑬(𝑸𝒕 ) = 𝝁, 𝒆𝒙𝒊𝒔𝒕𝒆 𝒆𝒕 𝒇𝒊𝒏𝒊𝒆 (𝟑)
𝑺𝒕−𝟏

̅ ↝ 𝓝(𝑬(𝑸
(𝟏) + (𝟐) + (𝟑) ⇒ (𝑸𝒕 )𝒕∈ℕ∗ 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒊. 𝒊. 𝒅. 𝒅𝒆 𝓝(𝝁, 𝝈𝟐 ) ⇒ 𝑸 ̅ ), 𝑽(𝑸
̅ ))
𝒏 𝒏 𝒏 𝒏
𝟏 𝟏 𝟏 𝟏
̅ ) = 𝑬 ( ∑ 𝑸𝒕 ) = 𝑬 (∑ 𝑸𝒕 ) = ∑ 𝑬(𝑸
𝑶ù 𝑬(𝑸 ⏟ 𝒕) = ∑ 𝝁 = 𝝁
𝒏 𝒏 𝒏 𝒏
𝒕=𝟏 𝒕=𝟏 𝒕=𝟏 𝝁 𝒕=𝟏
𝒏 𝒏 𝒏 𝒏
𝟏 𝟏 𝟏 𝟏 𝝈𝟐
̅ )
𝑬𝒕 𝑽(𝑸 = 𝑽 ( ∑ 𝑸𝒕 ) = 𝟐 𝑽 (∑ 𝑸𝒕 ) = 𝟐 ∑ 𝑽(𝑸 ) 𝟐
⏟ 𝒕 = 𝟐∑𝝈 =
𝒏 𝒏 𝒏 𝟐
𝒏 𝒏
𝒕=𝟏 𝒕=𝟏 𝒕=𝟏 𝝈 𝒕=𝟏

̅ ↝ 𝓝(𝝁, 𝝈𝟐 ⁄𝒏)
𝑪𝒆 𝒒𝒖𝒊 𝒅𝒐𝒏𝒏𝒆 ∶ 𝑸
̅ ̅
̅
𝑸
𝑬(𝑹𝟎,𝒏 ) = 𝑬(𝒆𝑸 − 𝟏) = 𝑬(𝒆𝑸 ) − 𝟏 = 𝑬(𝑯) − 𝟏
𝑶𝒏 𝒂 𝒅é𝒎𝒐𝒏𝒕𝒓é 𝒒𝒖𝒆 ∶ 𝑹𝟎,𝒏 = 𝒆 − 𝟏𝒅𝒐𝒏𝒄 , { ̅ ̅
𝑽(𝑹𝟎,𝒏 ) = 𝑽(𝒆𝑸 − 𝟏) = 𝑽(𝒆𝑸 ) = 𝑽(𝑯)
̅
𝒐ù 𝑯 = 𝒆𝑸
𝝈𝟐
(𝝁+ )
𝒀 ↝ 𝓝(𝝁, 𝝈𝟐 ) 𝟐
𝑪𝒐𝒎𝒑𝒕𝒆 𝒕𝒆𝒏𝒖 𝒅𝒆𝒔 𝒓é𝒔𝒖𝒍𝒕𝒂𝒕𝒔 𝒑𝒓é𝒄é𝒅𝒆𝒏𝒕𝒔 ∶ { 𝒀
⇒ { 𝑬(𝑿) = 𝒆
𝑿=𝒆 𝟐 𝟐
𝑽(𝑿) = 𝒆𝟐(𝝈 +𝝁) (𝟏 − 𝒆−𝝈 )

𝝈 𝟐
( )
𝟐 𝝁+ √𝒏
𝝈 𝟐
̅ ↝ 𝓝 (𝝁, (
𝑸 ) )
𝑰𝒍 𝒆𝒏 𝒓é𝒔𝒖𝒍𝒕𝒆 𝒒𝒖𝒆 ∶ { √𝒏 ⇒ 𝑬(𝑯) = 𝒆( )
𝟐
̅ 𝝈 𝝈 𝟐
𝑯 = 𝒆𝑸 𝟐(( ) +𝝁)
𝒏
−( )
𝑽(𝑯) = 𝒆 √ (𝟏 − 𝒆 √𝒏 )
{

𝝈𝟐 𝝈𝟐
(𝝁+ ) (𝝁+ )
𝟐𝒏 𝟐𝒏
𝑬(𝑯) = 𝒆 𝑬(𝑹𝟎,𝒏 ) = 𝒆 −𝟏
⇒ 𝝈𝟐 𝝈𝟐
𝑫′ 𝒐ù 𝝈𝟐
𝟐( +𝝁) 𝟐( +𝝁) 𝝈𝟐
𝑽(𝑯) = 𝒆 𝒏
(𝟏 − 𝒆− 𝒏 ) 𝑽(𝑹𝟎,𝒏 ) = 𝒆 𝒏
(𝟏 − 𝒆 −
𝒏 )
{ {

346 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
4)
𝝈𝟐
(𝝁+ )
𝟐𝒏
𝐥𝐢𝐦 𝑬(𝑹𝟎,𝒏 ) = 𝐥𝐢𝐦 [𝒆 − 𝟏] = 𝒆𝝁 − 𝟏
𝒏→+∞ 𝒏→+∞
𝐥𝐢𝐦 𝑬(𝑹𝟎,𝒏 ) = 𝒆𝝁 − 𝟏
′ 𝒏→+∞
𝝈𝟐
𝑫 𝒐ù {
𝟐(
𝒏
+𝝁) −
𝝈𝟐 𝐥𝐢𝐦 𝑽(𝑹𝟎,𝒏 ) = 𝟎
𝐥𝐢𝐦 𝑽(𝑹𝟎,𝒏 ) = 𝐥𝐢𝐦 [𝒆
⏟ (𝟏 − 𝒆 𝒏 )] = 𝟎 𝒏→+∞
𝒏→+∞ 𝒏→+∞ ⏟
𝒆𝟐𝝁
{ 𝟎

𝒎𝒒 𝒑
⇔ 𝑹𝟎,𝒏 → 𝒆𝝁 − 𝟏 ⇒ 𝑹𝟎,𝒏 → 𝒆𝝁 − 𝟏
𝒏→+∞ 𝒏→+∞

Exercice 22 : (Minimum de variables aléatoires continues et indépendantes-Modes de convergence)

ÉNONCÉ
𝑺𝒐𝒊𝒕 𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 𝒖𝒏𝒆 𝒔é𝒒𝒖𝒆𝒏𝒄𝒆 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒊. 𝒊. 𝒅 𝒑𝒂𝒓 𝒍𝒂 𝒍𝒐𝒊 ∶
−(𝒕−𝟐)
𝒇(𝒕) = {𝒆 , 𝒔𝒊 𝒕 ≥ 𝟐
𝟎, 𝒔𝒊 𝒏𝒐𝒏
𝑺𝒐𝒊𝒕 𝒀𝒏 = 𝐦𝐢𝐧 (𝑿𝒊 )
𝟏≤𝒊≤𝒏

1) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝒀𝒏
2) 𝑬𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒆𝒕 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒈é𝒏é𝒓𝒂𝒕𝒓𝒊𝒄𝒆 𝒅𝒆𝒔 𝒎𝒐𝒎𝒆𝒏𝒕𝒔 𝒅𝒆 𝒀𝒏
3) 𝑴𝒐𝒏𝒕𝒓𝒆𝒓 𝒒𝒖𝒆 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂 (𝒀𝒏 )𝒏∈ℕ 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒗𝒆𝒓𝒔 𝒖𝒏𝒆 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆 𝒒𝒖𝒆
𝒍’𝒐𝒏 𝒅é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓𝒂
𝟏
4) 𝑺𝒐𝒊𝒕 𝑻𝒏 = ∑𝒏𝒊=𝟏(𝑿𝒊 − 𝒀𝒏 ) .
𝒏
𝑴𝒐𝒏𝒕𝒓𝒆𝒓 𝒒𝒖𝒆 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂 (𝑻𝒏 )𝒏∈ℕ 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒗𝒆𝒓𝒔 𝒖𝒏𝒆 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕𝒆 𝒆𝒕
𝒅é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒔𝒂 𝒗𝒂𝒍𝒆𝒖𝒓.

Corrigé
1) 𝑬𝒕𝒂𝒏𝒕 𝒅𝒐𝒏𝒏é (𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 ) 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒆𝒕 𝒅𝒆 𝒎ê𝒎𝒆 𝒍𝒐𝒊 ; 𝒐𝒏 𝒂 ∶

∙ 𝑭𝒀𝒏 (𝒚) = 𝑷(𝒀𝒏 ≤ 𝒚) = 𝑷 (𝐦𝐢𝐧 (𝑿𝒊 ) ≤ 𝒚) = 𝟏 − 𝑷 (𝐦𝐢𝐧 (𝑿𝒊 ) > 𝒚)

𝟏≤𝒊≤𝒏 𝟏≤𝒊≤𝒏
𝒏 𝒏

= 𝟏 − 𝑷(𝑿𝟏 > 𝒚, 𝑿𝟐 > 𝒚 , … , 𝑿𝒏 > 𝒚) = 𝟏 − ∏ 𝑷(𝑿𝒊 > 𝒚) = 𝟏 − ∏[𝟏 − 𝑷(𝑿𝒊 ≤ 𝒚)]

𝒊=𝟏 𝒊=𝟏
𝒏 𝒏

= 𝟏 − ∏ (𝟏 − 𝑭𝑿𝒊 (𝒚)) = 𝟏 − ∏(𝟏 − 𝑭𝑿 (𝒚)) = 𝟏 − [𝟏 − 𝑭𝑿 (𝒚)]𝒏

𝒊=𝟏 𝒊=𝟏

𝑫′ 𝒐ù 𝑭𝒀𝒏 (𝒚) = 𝟏 − [𝟏 − 𝑭𝑿 (𝒚)]𝒏

∙ 𝑳𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝑻 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏 ∶

347 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒏 𝒏
𝒅[𝟏 − (𝟏 − 𝑭𝑿 (𝒚)) ] 𝒅[(𝟏 − 𝑭𝑿 (𝒚)) ] ′ 𝒏−𝟏
∀𝒚 ≥ 𝟐 ; 𝒇𝒀𝒏 (𝒚) = =− = −𝒏(𝟏 − 𝑭𝑿 (𝒚)) (𝟏 − 𝑭𝑿 (𝒚))
𝒅𝒚 𝒅𝒚
𝒏−𝟏
= 𝒏𝒇𝑿 (𝒚)(𝟏 − 𝑭𝑿 (𝒚))
𝟎 , 𝒔𝒊 𝒙 < 𝟐
−(𝒕−𝟐)
𝑶𝒓 𝒐𝒏 𝒂 ∶ 𝒇𝑿 (𝒕) = { 𝒆 , 𝒔𝒊 𝒕 ≥ 𝟐 ⇒ 𝑭𝑿 (𝒙) = {
𝒙
𝟎 , 𝒔𝒊 𝒏𝒐𝒏 ∫ 𝒆−(𝒕−𝟐) 𝒅𝒕 , 𝒔𝒊 𝒙 ≥ 𝟐
𝟐

𝟎 , 𝒔𝒊 𝒙 < 𝟐 𝟎 , 𝒔𝒊 𝒙 < 𝟐 𝟎 , 𝒔𝒊 𝒙 < 𝟐

𝑭𝑿 (𝒙) = { −(𝒕−𝟐) 𝒙 ⇒ 𝑭𝑿 (𝒙) = { −(𝒙−𝟐) ⇒ 𝑭𝑿 (𝒕) = {
−[𝒆 ]𝟐 , 𝒔𝒊 𝒙 ≥ 𝟐 −[𝒆 − 𝟏] , 𝒔𝒊 𝒙 ≥ 𝟐 𝟏 − 𝒆−(𝒙−𝟐) , 𝒔𝒊 𝒙 ≥ 𝟐
𝒏−𝟏 𝒏−𝟏
𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆, ∀𝒚 ≥ 𝟐 ; 𝒇𝒀𝒏 (𝒚) = 𝒏𝒇𝑿 (𝒚)(𝟏 − 𝑭𝑿 (𝒚)) = 𝒏𝒆−(𝒚−𝟐) [𝟏 − (𝟏 − 𝒆−(𝒚−𝟐) )]

∀𝒚 ≥ 𝟐 ; 𝒇𝒀𝒏 (𝒚) = 𝒏𝒆−(𝒚−𝟐) 𝒆−(𝒏−𝟏)(𝒚−𝟐) = 𝒏𝒆−𝒏(𝒚−𝟐)

𝒏𝒆−𝒏(𝒚−𝟐) , 𝒔𝒊 𝒚 ≥ 𝟐
𝑫′ 𝒐ù, 𝒇𝒀𝒏 (𝒚) = {
𝟎 , 𝒔𝒊 𝒏𝒐𝒏

2)
𝒏
∙ ∀𝒚 ≥ 𝟐, 𝑭𝒀𝒏 (𝒚) = 𝟏 − [𝟏 − 𝑭𝑿 (𝒚)]𝒏 = 𝟏 − [𝟏 − (𝟏 − 𝒆−(𝒚−𝟐) )] = 𝟏 − 𝒆−𝒏(𝒚−𝟐)

𝟎 , 𝒔𝒊 𝒚 < 𝟐
𝑭𝒀𝒏 (𝒚) = {
𝟏 − 𝒆−𝒏(𝒚−𝟐) , 𝒔𝒊 𝒚 ≥ 𝟐
+∞ +∞ +∞
∙ ∀𝒚 ≥ 𝟐, 𝑴𝒀𝒏 (𝒕) = 𝑬(𝒆 𝒕𝒀𝒏 )
=∫ 𝒕𝒚
𝒆 𝒇𝒀𝒏 (𝒚)𝒅𝒚 = ∫ 𝒏𝒆 𝒆 𝒕𝒚 −𝒏(𝒚−𝟐)
𝒅𝒚 = 𝒏 ∫ 𝒆[(𝒕−𝒏)𝒚+𝟐𝒏] 𝒅𝒚
𝟐 𝟐 𝟐
+∞
𝑶𝒓 𝑴𝒀𝒏 (𝒕) 𝒆𝒙𝒊𝒔𝒕𝒆 𝒔𝒊 𝒆𝒕 𝒔𝒆𝒖𝒍𝒆𝒎𝒆𝒏𝒕 𝒔𝒊 ∫ 𝒆[(𝒕−𝒏)𝒚+𝟐𝒏] 𝒅𝒚 𝒄‑à‑𝒅. ∶ 𝒕 − 𝒏 < 𝟎 𝒐𝒖 𝒆𝒏𝒄𝒐𝒓𝒆 𝒕 ∈ ]−∞, 𝒏[
𝟐

+∞
𝒆[(𝒕−𝒏)𝒚+𝟐𝒏] 𝒏
∀𝒕 ∈ ]−∞, 𝒏[ , 𝑴𝒀𝒏 (𝒕) = 𝒏 [ ] = [( 𝐥𝐢𝐦 𝒆[(𝒕−𝒏)𝒚+𝟐𝒏] ) − (𝒆[𝟐(𝒕−𝒏)+𝟐𝒏] )]
𝒕−𝒏 𝟐
𝒕−𝒏 ⏟𝒚→+∞
𝟎

−𝒏 [𝟐𝒕−𝟐𝒏+𝟐𝒏] 𝒏𝒆𝟐𝒕
= (𝒆 )=
𝒕−𝒏 𝒏−𝒕
𝒏𝒆𝟐𝒕
∀𝒕 ∈ ]−∞, 𝒏[ , 𝑴𝒀𝒏 (𝒕) =
𝒏−𝒕
3)
∙ 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝑬(𝒀𝒏 ) 𝒆𝒕 𝑽(𝒀𝒏 ) à 𝒑𝒂𝒓𝒕𝒊𝒓 𝒅𝒆 𝑴𝒀𝒏 (𝒕) ∶
′
𝒅𝑴𝒀𝒏 (𝒕) 𝒆𝟐𝒕 (𝒆𝟐𝒕 )′ (𝒏 − 𝒕) − 𝒆𝟐𝒕 (𝒏 − 𝒕)′ 𝟐𝒆𝟐𝒕 (𝒏 − 𝒕) + 𝒆𝟐𝒕
= 𝒏( ) = 𝒏( ) = 𝒏( )
𝒅𝒕 𝒏−𝒕 (𝒏 − 𝒕)𝟐 (𝒏 − 𝒕)𝟐

𝒅𝑴𝒀𝒏 (𝒕) 𝒏𝒆𝟐𝒕 (𝟏 + 𝟐(𝒏 − 𝒕)) ′ (𝟎)

𝒏𝒆𝟎 (𝟏 + 𝟐(𝒏 − 𝟎)) 𝒏(𝟐𝒏 + 𝟏) 𝟐𝒏 + 𝟏
= ⇒ 𝑬(𝒀𝒏 ) = 𝑴 𝒀𝒏 = = =
𝒅𝒕 (𝒏 − 𝒕)𝟐 (𝒏 − 𝟎)𝟐 𝒏𝟐 𝒏

348 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝟐𝒏 + 𝟏
𝑬(𝒀𝒏 ) =
𝒏
′ ′
𝒅𝟐 𝑴𝒀𝒏 (𝒕) 𝒏𝒆𝟐𝒕 (𝟏 + 𝟐(𝒏 − 𝒕)) [𝒆𝟐𝒕 (𝟏 + 𝟐(𝒏 − 𝒕))] (𝒏 − 𝒕)𝟐 − 𝒆𝟐𝒕 (𝟏 + 𝟐(𝒏 − 𝒕))[(𝒏 − 𝒕)𝟐 ]′
=[ ] = 𝒏[ ]
𝒅𝒕𝟐 (𝒏 − 𝒕)𝟐 (𝒏 − 𝒕)𝟒
′ ′
▪[𝒆𝟐𝒕 (𝟏 + 𝟐(𝒏 − 𝒕))] = (𝒆𝟐𝒕 )′ (𝟏 + 𝟐(𝒏 − 𝒕)) + 𝒆𝟐𝒕 (𝟏 + 𝟐(𝒏 − 𝒕)) = 𝟐𝒆𝟐𝒕 (𝟏 + 𝟐(𝒏 − 𝒕)) − 𝟐𝒆𝟐𝒕

= 𝟐𝒆𝟐𝒕 (𝟏 + 𝟐(𝒏 − 𝒕) − 𝟏) = 𝟒𝒆𝟐𝒕 (𝒏 − 𝒕)

▪[(𝒏 − 𝒕)𝟐 ]′ = 𝟐(𝒏 − 𝒕)′ (𝒏 − 𝒕) = −𝟐(𝒏 − 𝒕)
𝒅𝟐 𝑴𝒀𝒏 (𝒕) 𝟒𝒆𝟐𝒕 (𝒏 − 𝒕)𝟑 + 𝟐𝒆𝟐𝒕 (𝟏 + 𝟐(𝒏 − 𝒕))(𝒏 − 𝒕) 𝟐𝒏𝒆𝟐𝒕 [𝟐(𝒏 − 𝒕)𝟐 + 𝟐(𝒏 − 𝒕) + 𝟏]
= 𝒏[ ]=
𝒅𝒕𝟐 (𝒏 − 𝒕)𝟒 (𝒏 − 𝒕)𝟑

𝒅𝟐 𝑴𝒀𝒏 (𝒕) 𝟐𝒏𝒆𝟐𝒕 [𝟐(𝒏 − 𝒕)𝟐 + 𝟐(𝒏 − 𝒕) + 𝟏]

=
𝒅𝒕𝟐 (𝒏 − 𝒕)𝟑
𝟐𝒏𝒆𝟎 [𝟐(𝒏 − 𝟎)𝟐 + 𝟐(𝒏 − 𝟎) + 𝟏] 𝟐𝒏[𝟐𝒏𝟐 + 𝟐𝒏 + 𝟏] 𝟐(𝟐𝒏𝟐 + 𝟐𝒏 + 𝟏)
⇒ 𝑬(𝒀𝟐𝒏 ) = 𝑴′′
𝒀𝒏 (𝟎) = = =
(𝒏 − 𝟎)𝟑 𝒏𝟑 𝒏𝟐
𝟐(𝟐𝒏𝟐 + 𝟐𝒏 + 𝟏) (𝟐𝒏 + 𝟏)𝟐 𝟒𝒏𝟐 + 𝟒𝒏 + 𝟐 − 𝟒𝒏𝟐 − 𝟒𝒏 − 𝟏 𝟏
𝑽(𝒀𝒏 ) = 𝑬(𝒀𝟐𝒏 ) − [𝑬(𝒀𝒏 )]𝟐 = 𝟐
− 𝟐
= 𝟐
= 𝟐
𝒏 𝒏 𝒏 𝒏
𝟏
𝑽(𝒀𝒏 ) =
𝒏𝟐
𝟐𝒏 + 𝟏
𝐥𝐢𝐦 𝑬(𝒀𝒏 ) = 𝐥𝐢𝐦 =𝟐 𝒎𝒒 𝒑
∙{
𝒏→+∞ 𝒏→+∞ 𝒏 ⇔ 𝒀𝒏 → 𝟐 ⇒ 𝒀𝒏 → 𝟐
𝟏 𝒏→+∞ 𝒏→+∞
𝐥𝐢𝐦 𝑽(𝒀𝒏 ) = 𝐥𝐢𝐦 𝟐 = 𝟎
𝒏→+∞ 𝒏→+∞ 𝒏

4)
[𝟐, +∞[
∙ 𝑺𝒐𝒊𝒕 𝒁 = 𝑿 − 𝟐 , 𝒂𝒊𝒏𝒔𝒊 {𝑿(𝛀) = ⇒ 𝒁(𝛀) = [𝟎, +∞[
𝒁 =𝑿−𝟐
𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒐𝒏𝒔 𝒍𝒂 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒓é𝒑𝒂𝒓𝒕𝒊𝒕𝒊𝒐𝒏 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒁 :
𝑭𝒁 (𝒛) = 𝑷(𝒁 ≤ 𝒛) = 𝑷(𝑿 − 𝟐 ≤ 𝒛) = 𝑷(𝑿 ≤ 𝒛 + 𝟐) = 𝑭𝑿 (𝒛 + 𝟐)
𝑳𝒂 𝒅. 𝒅. 𝒑 𝒅𝒆 𝒍𝒂 𝒗. 𝒂. 𝒁 𝒔𝒆 𝒅é𝒅𝒖𝒊𝒕 𝒅𝒆 𝒔𝒂 𝒇. 𝒓 𝒑𝒂𝒓 𝒅é𝒓𝒊𝒗𝒂𝒕𝒊𝒐𝒏 ∶
𝒅[𝑭𝒁 (𝒛)]
𝒇𝒁 (𝒛) = = (𝒛 + 𝟐)′ 𝒇𝑿 (𝒛 + 𝟐) = 𝒆−((𝒛+𝟐)−𝟐) = 𝒆−𝒛
𝒅𝒛
𝒆−𝒛 , 𝒔𝒊 𝒛 ≥ 𝟎 𝑬(𝑿 − 𝟐) = 𝟏 𝑬(𝑿) = 𝟑
𝒇𝒁 (𝒛) = { ⇒ 𝒁 ↝ 𝓔(𝟏) ⇒ 𝑬(𝒁) = 𝑽(𝒁) = 𝟏 ⇒ { ⇒{
𝟎 , 𝒔𝒊 𝒏𝒐𝒏 𝑽(𝑿 − 𝟐) = 𝟏 𝑽(𝑿) = 𝟏
𝒏 𝒏 𝒏
𝟏 𝟏 𝟏
̅−𝒀
∙ 𝑶𝒏 𝒂 𝑻𝒏 = ∑(𝑿𝒊 − 𝒀𝒏 ) = ∑ 𝑿𝒊 − ∑ 𝒀𝒏 = 𝑿 ̅𝒏
𝒏 𝒏 𝒏
𝒊=𝟏 𝒊=𝟏 𝒊=𝟏

349 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒏
𝒑 𝟏 𝒑
▪𝒀𝒏 → 𝟐⇒𝒀 ̅ 𝒏 = ∑ 𝒀𝒏 → 𝟐
𝒏→+∞ 𝒏 𝒏→+∞
𝒊=𝟏
𝒏 𝒏 𝒏
𝟏 𝟏 𝟏
̅ ) = 𝐥𝐢𝐦 𝑬 ( ∑ 𝑿𝒊 ) = 𝐥𝐢𝐦 𝑬 (∑ 𝑿𝒊 ) = 𝐥𝐢𝐦 ∑ 𝑬
𝐥𝐢𝐦 𝑬(𝑿 ⏟(𝑿𝒊 ) = 𝟑
𝒏→+∞ 𝒏→+∞ 𝒏 𝒏→+∞ 𝒏 𝒏→+∞ 𝒏
𝒊=𝟏 𝒊=𝟏 𝒊=𝟏 𝟑
▪ 𝒏 𝒏 𝒏
𝟏 𝟏 𝟏
̅ ) = 𝐥𝐢𝐦 𝑽 ( ∑ 𝑿𝒊 ) = 𝐥𝐢𝐦 𝟐 𝑽 (∑ 𝑿𝒊 ) = 𝐥𝐢𝐦 𝟐 [∑ ⏟
𝐥𝐢𝐦 𝑽(𝑿 𝑽𝒂𝒓(𝑿𝒊 ) + 𝟐 ∑ ⏟
𝑪𝒐𝒗(𝑿𝒊 , 𝑿𝒋 )]
𝒏→+∞ 𝒏→+∞ 𝒏 𝒏→+∞ 𝒏 𝒏→+∞ 𝒏
{ 𝒊=𝟏 𝒊=𝟏 𝒊=𝟏 𝟏 𝟏≤𝒊<𝑗≤𝑛 𝟎

𝟏 𝒎𝒒 𝒑
̅ ) = 𝟑 𝒆𝒕 𝐥𝐢𝐦 𝑽(𝑿
( 𝐥𝐢𝐦 𝑬(𝑿 ̅ ) = 𝐥𝐢𝐦 ̅→
= 𝟎) ⇔ 𝑿 ̅→
𝟑⇒𝑿 𝟑
𝒏→+∞ 𝒏→+∞ 𝒏→+∞ 𝒏 𝒏→+∞ 𝒏→+∞
𝒑
̅𝒏 →
𝒀 𝟐 𝒑
𝒏→+∞ ̅−𝒀
̅𝒏 →
▪{ 𝒑 ⇒ 𝑻𝒏 = 𝑿 𝟏
̅→ 𝒏→+∞
𝑿 𝟑
𝒏→+∞

Exercice 23 : (Variables de Bernoulli-Modes de convergence)

ÉNONCÉ
𝑺𝒐𝒊𝒕 (𝑿𝒏 )𝒏∈ℕ∗ 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝒅𝒆 𝑩𝒆𝒓𝒏𝒐𝒖𝒍𝒍𝒊, 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔, 𝒅𝒆 𝒎ê𝒎𝒆
𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆 𝒑 ∈ ]𝟎, 𝟏[ . 𝑷𝒐𝒖𝒓 𝒕𝒐𝒖𝒕 𝒆𝒏𝒕𝒊𝒆𝒓 𝒏𝒂𝒕𝒖𝒓𝒆𝒍 𝒏 𝒏𝒐𝒏 𝒏𝒖𝒍, 𝒐𝒏 𝒅é𝒇𝒊𝒏𝒊𝒕 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆
𝒀𝒏 = 𝟐𝑿𝒏 + 𝑿𝒏+𝟏 − 𝑿𝒏 𝑿𝒏+𝟏
1) 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝒀𝒏 , 𝒄𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒔𝒐𝒏 𝒆𝒔𝒑é𝒓𝒂𝒏𝒄𝒆 𝒆𝒕 𝒔𝒂 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒆𝒏 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝒑.
2)
𝟏
a) 𝑶𝒏 𝒏𝒐𝒕𝒆 𝒁𝒏 = ∑𝒏𝒌=𝟏 𝒀𝒌 , 𝒄𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝑬(𝒁𝒏 )
𝒏
b) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝑪𝒐𝒗(𝒀𝒊 , 𝒀𝒋 ) , 𝒑𝒐𝒖𝒓 𝟏 ≤ 𝒊 < 𝒋 ≤ 𝒏
c) 𝑽𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝑽(𝒁𝒏 )
3) 𝑴𝒐𝒏𝒕𝒓𝒆𝒓 𝒒𝒖𝒆 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 (𝒁𝒏 )𝒏∈ℕ∗ , 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒗𝒆𝒓𝒔 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒄𝒆𝒓𝒕𝒂𝒊𝒏𝒆
é𝒈𝒂𝒍𝒆 à 𝟑𝒑 − 𝒑𝟐

Corrigé
1) 𝑿𝒏 ↝ 𝓑 (𝟏, 𝒑 ) ⇒ 𝑿𝒏 (𝛀) = 𝑿𝒏 (𝛀) = {𝟎, 𝟏}
𝑶𝒏 𝒂 ∶ 𝒀𝒏 = 𝟐𝑿𝒏 + 𝑿𝒏+𝟏 − 𝑿𝒏 𝑿𝒏+𝟏 , 𝒅𝒐𝒏𝒄 ∶
𝑿𝒏+𝟏 𝑿𝒏+𝟏 = 𝟎 𝑿𝒏+𝟏 = 𝟏
𝑿𝒏
𝑿𝒏 = 𝟎 𝒀𝒏 = 𝟐𝑿𝒏 + 𝑿𝒏+𝟏 − 𝑿𝒏 𝑿𝒏+𝟏 = 𝟎 𝒀𝒏 = 𝟐𝑿𝒏 + 𝑿𝒏+𝟏 − 𝑿𝒏 𝑿𝒏+𝟏 = 𝟏 ⇒ 𝒀𝒏 (𝛀) = {𝟎, 𝟏, 𝟐}
𝑿𝒏 = 𝟏 𝒀𝒏 = 𝟐𝑿𝒏 + 𝑿𝒏+𝟏 − 𝑿𝒏 𝑿𝒏+𝟏 = 𝟐 𝒀𝒏 = 𝟐𝑿𝒏 + 𝑿𝒏+𝟏 − 𝑿𝒏 𝑿𝒏+𝟏 = 𝟐
∙ 𝑷(𝒀𝒏 = 𝟎) = 𝑷(𝑿𝒏 = 𝟎, 𝑿𝒏+𝟏 = 𝟎) = 𝑷(𝑿𝒏 = 𝟎)𝑷(𝑿𝒏+𝟏 = 𝟎) = (𝟏 − 𝒑)𝟐
∙ 𝑷(𝒀𝒏 = 𝟏) = 𝑷(𝑿𝒏 = 𝟎, 𝑿𝒏+𝟏 = 𝟏) = 𝑷(𝑿𝒏 = 𝟎)𝑷(𝑿𝒏+𝟏 = 𝟏) = 𝒑(𝟏 − 𝒑)
∙ 𝑷(𝒀𝒏 = 𝟐) = 𝑷(𝑿𝒏 = 𝟏, 𝑿𝒏+𝟏 = 𝟎) + 𝑷(𝑿𝒏 = 𝟏, 𝑿𝒏+𝟏 = 𝟏)

350 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
= 𝑷(𝑿𝒏 = 𝟏)𝑷(𝑿𝒏+𝟏 = 𝟎) + 𝑷(𝑿𝒏 = 𝟏)𝑷(𝑿𝒏+𝟏 = 𝟏) = 𝒑(𝟏 − 𝒑) + 𝒑𝟐 = 𝒑

∙ 𝑬(𝒀𝒏 ) = ∑ 𝒚𝒏 𝑷(𝒀𝒏 = 𝒚𝒏 ) = [𝟎 × 𝑷(𝒀𝒏 = 𝟎)] + [𝟏 × 𝑷(𝒀𝒏 = 𝟏)] + [𝟐 × 𝑷(𝒀𝒏 = 𝟐)]

𝒚𝒏 ∈{𝟎,𝟏,𝟐}

= 𝒑(𝟏 − 𝒑) + 𝟐𝒑 = 𝟑𝒑 − 𝒑𝟐 = 𝒑(𝟑 − 𝒑) . 𝑫′ 𝒐ù 𝑬(𝒀𝒏 ) = 𝒑(𝟑 − 𝒑)

∙ 𝑬(𝒀𝟐𝒏 ) = ∑ 𝒚𝟐𝒏 𝑷(𝒀𝒏 = 𝒚𝒏 ) = 𝟎 + 𝑷(𝒀𝒏 = 𝟏) + 𝟒 × 𝑷(𝒀𝒏 = 𝟐) = 𝒑(𝟏 − 𝒑) + 𝟒𝒑 = 𝟓𝒑 − 𝒑𝟐

𝒚𝒏 ∈{𝟎,𝟏,𝟐}

∙ 𝑽(𝒀𝒏 ) = 𝑬(𝒀𝟐𝒏 ) − [𝑬(𝒀𝒏 )]𝟐 = 𝟓𝒑 − 𝒑𝟐 − [𝒑(𝟑 − 𝒑)]𝟐 = 𝟓𝒑 − 𝒑𝟐 − 𝒑𝟐 (𝟑 − 𝒑)𝟐

= 𝒑[𝟓 − 𝒑 − 𝒑(𝟑 − 𝒑)𝟐 ] = 𝒑[𝟓 − 𝒑 − 𝒑(𝟗 − 𝟔𝒑 + 𝒑𝟐 )] = 𝒑[𝟓 − 𝒑 − (𝟗𝒑 − 𝟔𝒑𝟐 + 𝒑𝟑 )]
= 𝒑[𝟓 − 𝒑 − 𝟗𝒑 + 𝟔𝒑𝟐 − 𝒑𝟑 ] = 𝒑(𝟏 − 𝒑)(𝒑𝟐 − 𝟓𝒑 + 𝟓).

𝑫′ 𝒐ù 𝑽(𝒀𝒏 ) = 𝒑(𝟏 − 𝒑)(𝒑𝟐 − 𝟓𝒑 + 𝟓)

2)
a)
𝒏 𝒏 𝒏 𝒏
𝟏 𝟏 𝟏 𝟏
𝑬(𝒁𝒏 ) = 𝑬 ( ∑ 𝒀𝒌 ) = 𝑬 (∑ 𝒀𝒌 ) = ∑ 𝑬(𝒀𝒌 ) = ∑[𝒑(𝟑 − 𝒑)] . 𝑫′ 𝒐ù 𝑬(𝒁𝒏 ) = 𝒑(𝟑 − 𝒑)
𝒏 𝒏 𝒏 𝒏
𝒌=𝟏 𝒌=𝟏 𝒌=𝟏 𝒌=𝟏

b)
∙ 𝑺𝒊 𝒊 + 𝟏 ≠ 𝒋 𝒐𝒖 𝒆𝒏𝒄𝒐𝒓𝒆 𝒔𝒊 𝒊 + 𝟏 < 𝒋 ⇒ 𝒊 ≠ 𝒋 , 𝒊 ≠ 𝒋 + 𝟏 𝒆𝒕 𝒊 + 𝟏 ≠ 𝒋 + 𝟏

𝒀𝒊 𝒀𝒋 = (𝟐𝑿𝒊 + 𝑿𝒊+𝟏 − 𝑿𝒊 𝑿𝒊+𝟏 )( 𝟐𝑿𝒋 + 𝑿𝒋+𝟏 − 𝑿𝒋 𝑿𝒋+𝟏 )

= 𝟒𝑿𝒊 𝑿𝒋 + 𝟐𝑿𝒊 𝑿𝒋+𝟏 − 𝟐𝑿𝒊 𝑿𝒋 𝑿𝒋+𝟏 + 𝟐𝑿𝒊+𝟏 𝑿𝒋 + 𝑿𝒊+𝟏 𝑿𝒋+𝟏 − 𝑿𝒊+𝟏 𝑿𝒋 𝑿𝒋+𝟏 − 𝟐𝑿𝒊 𝑿𝒊+𝟏 𝑿𝒋 − 𝑿𝒊 𝑿𝒊+𝟏 𝑿𝒋+𝟏 + 𝑿𝒊 𝑿𝒊+𝟏 𝑿𝒋 𝑿𝒋+𝟏
𝒏 𝒏 𝒏

𝑶𝒓 (𝑿𝒏 )𝒏∈ℕ∗ 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒊. 𝒊. 𝒅 𝒅𝒆 𝓑 (𝟏, 𝒑 ) ⇒ 𝑬 (∏ 𝑿𝒌 ) = ∏ 𝑬(𝑿𝒌 ) = ∏ 𝒑 = 𝒑𝒏

𝒌=𝟏 𝒌=𝟏 𝒌=𝟏

𝑨𝒊𝒏𝒔𝒊 𝑬(𝒀𝒊 𝒀𝒋 ) = 𝟒𝒑𝟐 + 𝟐𝒑𝟐 − 𝟐𝒑𝟑 + 𝟐𝒑𝟐 + 𝒑𝟐 − 𝒑𝟑 − 𝟐𝒑𝟑 − 𝒑𝟑 + 𝒑𝟒

𝑬(𝒀𝒊 𝒀𝒋 ) = 𝟗𝒑𝟐 − 𝟔𝒑𝟑 + 𝒑𝟒 = (𝟑𝒑 − 𝒑𝟐 )𝟐 = 𝒑𝟐 (𝟑 − 𝒑)𝟐

𝑪𝒐𝒗(𝒀𝒊 , 𝒀𝒋 ) = 𝑬(𝒀𝒊 𝒀𝒋 ) − 𝑬(𝒀𝒊 )𝑬(𝒀𝒋 ) = 𝒑𝟐 (𝟑 − 𝒑)𝟐 − 𝒑𝟐 (𝟑 − 𝒑)𝟐 = 𝟎

𝑫𝒐ù , 𝒔𝒊 𝒊 + 𝟏 ≠ 𝒋 ∶ 𝑪𝒐𝒗(𝒀𝒊 , 𝒀𝒋 ) = 𝟎

∙ 𝑺𝒊 𝒊 + 𝟏 = 𝒋
𝒀𝒊 𝒀𝒋 = 𝒀𝒊 𝒀𝒊+𝟏 = 𝟒𝑿𝒊 𝑿𝒊+𝟏 + 𝟐𝑿𝒊 𝑿𝒊+𝟐 − 𝟐𝑿𝒊 𝑿𝒊+𝟏 𝑿𝒊+𝟐 + 𝟐𝑿𝟐𝒊+𝟏 + 𝑿𝒊+𝟏 𝑿𝒊+𝟐 − 𝑿𝟐𝒊+𝟏 𝑿𝒊+𝟐 − 𝟐𝑿𝒊 𝑿𝟐𝒊+𝟏 − 𝑿𝒊 𝑿𝒊+𝟏 𝑿𝒊+𝟐 + 𝑿𝒊 𝑿𝟐𝒊+𝟏 𝑿𝒊+𝟐

𝑨𝒊𝒏𝒔𝒊 𝑬(𝒀𝒊 𝒀𝒋 ) = 𝟒𝒑𝟐 + 𝟐𝒑𝟐 − 𝟐𝒑𝟑 + 𝟐𝒑𝟐 + 𝟐𝒑 − 𝒑𝟐 − 𝒑𝟑 − 𝑬(𝑿𝟐𝒊+𝟏 𝑿𝒊+𝟐 ) − 𝟐𝑬(𝑿𝒊 𝑿𝟐𝒊+𝟏 ) + 𝑬(𝑿𝒊 𝑿𝟐𝒊+𝟏 𝑿𝒊+𝟐)

= 𝟐𝒑 + 𝟕𝒑𝟐 − 𝟑𝒑𝟑 − 𝑬(𝑿𝟐𝒊+𝟏 𝑿𝒊+𝟐 ) − 𝟐𝑬(𝑿𝒊 𝑿𝟐𝒊+𝟏 ) + 𝑬(𝑿𝒊 𝑿𝟐𝒊+𝟏 𝑿𝒊+𝟐)

𝑷(𝑿𝟐𝒌 = 𝟎) = 𝑷(𝑿𝒌 = 𝟎) = 𝟏 − 𝒑
▪𝑿𝒌 ↝ 𝓑 (𝟏, 𝒑 ) ⇒ 𝑿𝒌 (𝛀) = 𝑿𝟐𝒌 (𝛀) = {𝟎, 𝟏} 𝒆𝒕 { ⇒ 𝑿𝟐𝒌 ↝ 𝓑 (𝟏, 𝒑 )
𝑷(𝑿𝟐𝒌 = 𝟏) = 𝑷(𝑿𝒌 = 𝟏) = 𝒑

351 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝑷(𝑿𝟐𝒌+𝒍 = 𝟎, 𝑿𝒌 = 𝟎) = 𝑷(𝑿𝒌+𝒍 = 𝟎, 𝑿𝒌 = 𝟎) = 𝑷(𝑿𝒌+𝒍 = 𝟎)𝑷(𝑿𝒌 = 𝟎) = (𝟏 − 𝒑)𝟐 = 𝑷(𝑿𝟐𝒌+𝒍 = 𝟎)𝑷(𝑿𝒌 = 𝟎) , 𝒍 ∈ ℕ∗

𝑷(𝑿𝟐𝒌+𝒍 = 𝟏, 𝑿𝒌 = 𝟎) = 𝑷(𝑿𝒌+𝒍 = 𝟏, 𝑿𝒌 = 𝟎) = 𝑷(𝑿𝒌+𝒍 = 𝟏)𝑷(𝑿𝒌 = 𝟎) = 𝒑(𝟏 − 𝒑) = 𝑷(𝑿𝟐𝒌+𝒍 = 𝟏)𝑷(𝑿𝒌 = 𝟎), 𝒍 ∈ ℕ∗

𝑷(𝑿𝟐𝒌+𝒍 = 𝟎, 𝑿𝒌 = 𝟏) = 𝑷(𝑿𝒌+𝒍 = 𝟎, 𝑿𝒌 = 𝟏) = 𝑷(𝑿𝒌+𝒍 = 𝟎)𝑷(𝑿𝒌 = 𝟏) = 𝒑(𝟏 − 𝒑) = 𝑷(𝑿𝟐𝒌+𝒍 = 𝟎)𝑷(𝑿𝒌 = 𝟏), 𝒍 ∈ ℕ∗

𝑷(𝑿𝟐𝒌+𝒍 = 𝟏, 𝑿𝒌 = 𝟏) = 𝑷(𝑿𝒌+𝒍 = 𝟏, 𝑿𝒌 = 𝟏) = 𝑷(𝑿𝒌+𝒍 = 𝟏)𝑷(𝑿𝒌 = 𝟏) = 𝒑𝟐 = 𝑷(𝑿𝟐𝒌+𝒍 = 𝟏)𝑷(𝑿𝒌 = 𝟏), 𝒍 ∈ ℕ∗

𝑿𝟐𝒊+𝟏 𝒆𝒕 𝑿𝒊 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒊. 𝒊. 𝒅 𝒅𝒆 𝓑 (𝟏, 𝒑 )

𝑶𝒏 𝒐𝒃𝒕𝒊𝒆𝒏𝒕 ∶ ∀ 𝒍 ∈ ℕ∗ , 𝑿𝟐𝒌+𝒍 𝒆𝒕 𝑿𝒌 𝒅𝒆 𝒗. 𝒂. 𝒊. 𝒊. 𝒅 𝒅𝒆 𝓑 (𝟏, 𝒑 ) ⇒ {
𝑿𝟐𝒊+𝟏 𝒆𝒕 𝑿𝒊+𝟐 𝒅𝒆𝒖𝒙 𝒗. 𝒂. 𝒊. 𝒊. 𝒅 𝒅𝒆 𝓑 (𝟏, 𝒑 )

⇒ 𝑿𝒊 , 𝑿𝟐𝒊+𝟏 𝒆𝒕 𝑿𝒊+𝟐 𝒅𝒆𝒔 𝒗. 𝒂. 𝒊. 𝒊. 𝒅 𝒅𝒆 𝓑 (𝟏, 𝒑 )

𝑪𝒆 𝒒𝒖𝒊 𝒅𝒐𝒏𝒏𝒆 ∶ 𝑬(𝒀𝒊 𝒀𝒋 ) = 𝟐𝒑 + 𝟕𝒑𝟐 − 𝟑𝒑𝟑 − 𝑬(𝑿𝟐𝒊+𝟏 )𝑬(𝑿𝒊+𝟐 ) − 𝟐𝑬(𝑿𝒊 )𝑬(𝑿𝟐𝒊+𝟏 ) + 𝑬(𝑿𝒊 )𝑬(𝑿𝟐𝒊+𝟏 )𝑬(𝑿𝒊+𝟐 )

= 𝟐𝒑 + 𝟕𝒑𝟐 − 𝟑𝒑𝟑 − 𝒑𝟐 − 𝟐𝒑𝟐 + 𝒑𝟑 = 𝟐𝒑 + 𝟒𝒑𝟐 − 𝟐𝒑𝟑

𝑪𝒐𝒗(𝒀𝒊 , 𝒀𝒋 ) = 𝑬(𝒀𝒊 𝒀𝒋 ) − 𝑬(𝒀𝒊 )𝑬(𝒀𝒋 ) = 𝟐𝒑 + 𝟒𝒑𝟐 − 𝟐𝒑𝟑 − 𝒑𝟐 (𝟑 − 𝒑)𝟐

= 𝒑[(𝟐 + 𝟒𝒑 − 𝟐𝒑𝟐 ) − 𝒑(𝟑 − 𝒑)𝟐 ] = 𝒑[𝟐 + 𝟒𝒑 − 𝟐𝒑𝟐 − 𝒑(𝟗 − 𝟔𝒑 + 𝒑𝟐 )] = 𝒑[𝟐 + 𝟒𝒑 − 𝟐𝒑𝟐 − 𝟗𝒑 + 𝟔𝒑𝟐 − 𝒑𝟑 ]

= 𝒑[𝟐 − 𝟓𝒑 + 𝟒𝒑𝟐 − 𝒑𝟑 ] = 𝒑(𝟏 − 𝒑)(𝟐 − 𝟑𝒑 + 𝒑𝟐 ) = 𝒑(𝟏 − 𝒑)(𝟏 − 𝒑)(𝟐 − 𝒑) = 𝒑(𝒑 − 𝟏)𝟐 (𝟐 − 𝒑)

𝟎 , 𝒔𝒊 , 𝒊 + 𝟏 < 𝒋
𝑫′ 𝒐ù 𝑪𝒐𝒗(𝒀𝒊 , 𝒀𝒋 ) = {
𝒑(𝒑 − 𝟏)𝟐 (𝟐 − 𝒑) , 𝒔𝒊 , 𝒊 + 𝟏 = 𝒋

c)
𝒏 𝒏
𝟏 𝟏
𝑽(𝒁𝒏 ) = 𝑽 ( ∑ 𝒀𝒊 ) = 𝟐 [∑ 𝑽𝒂𝒓(𝒀𝒊 ) + 𝟐 ∑ 𝑪𝒐𝒗(𝒀𝒊 , 𝒀𝒋 )]
𝒏 𝒏
𝒊=𝟏 𝒊=𝟏 𝟏≤𝒊<𝑗≤𝑛

𝒏 𝒏−𝟏 𝒏
𝟏
= 𝟐 [∑ 𝒑(𝟏 − 𝒑)(𝒑𝟐 − 𝟓𝒑 + 𝟓) + 𝟐 ∑ ∑ 𝑪𝒐𝒗(𝒀𝒊 , 𝒀𝒋 )]
𝒏
𝒌=𝟏 𝒊=𝟏 𝒋=𝒊+𝟏

𝒏−𝟏 𝒏
𝟏
= 𝟐 𝒏𝒑(𝟏 − 𝒑)(𝒑𝟐 − 𝟓𝒑 + 𝟓) + 𝟐 ∑ 𝑪𝒐𝒗(𝒀𝒊 , 𝒀𝒊+𝟏 ) + ∑ 𝑪𝒐𝒗(𝒀𝒊 , 𝒀𝒋 )
𝒏
𝒊=𝟏 ⏟
𝒋=𝒊+𝟐
[ ( 𝟎 )]
𝒏−𝟏
𝟏
= 𝟐 [𝒏𝒑(𝟏 − 𝒑)(𝒑𝟐 − 𝟓𝒑 + 𝟓) + 𝟐 ∑ 𝒑(𝟏 − 𝒑)𝟐 (𝟐 − 𝒑)]
𝒏
𝒊=𝟏

𝒏𝒑(𝟏 − 𝒑)(𝒑𝟐 − 𝟓𝒑 + 𝟓) + 𝟐(𝒏 − 𝟏)𝒑(𝟏 − 𝒑)𝟐 (𝟐 − 𝒑) 𝒑(𝟏 − 𝒑) (𝒏(𝒑𝟐 − 𝟓𝒑 + 𝟓) + 𝟐(𝒏 − 𝟏)(𝟏 − 𝒑)(𝟐 − 𝒑))
= =
𝒏𝟐 𝒏𝟐

𝒑(𝟏 − 𝒑) (𝟓𝒏 − 𝟓𝒏𝒑 + 𝒏𝒑𝟐 + 𝟐(𝒏 − 𝟏)(𝟐 − 𝟑𝒑 + 𝒑𝟐 ))

=
𝒏𝟐

𝒑(𝟏 − 𝒑)(𝟓𝒏 − 𝟓𝒏𝒑 + 𝒏𝒑𝟐 + 𝟒(𝒏 − 𝟏) − 𝟔(𝒏 − 𝟏)𝒑 + 𝟐(𝒏 − 𝟏)𝒑𝟐 ) 𝒑(𝟏 − 𝒑)(𝟗𝒏 − 𝟒 − 𝟏𝟏𝒏𝒑 + 𝟔𝒑 + 𝟑𝒏𝒑𝟐 − 𝟐𝒑𝟐 )
= =
𝒏𝟐 𝒏𝟐

′
(𝒏(𝟗 − 𝟏𝟏𝒑 + 𝟑𝒑𝟐 ) − 𝟒 + 𝟔𝒑 − 𝟐𝒑𝟐 )𝒑(𝟏 − 𝒑)
𝑫 𝒐ù, 𝑽(𝒁𝒏 ) =
𝒏𝟐

352 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
3)
𝐥𝐢𝐦 𝑬(𝒁𝒏 ) = 𝒑(𝟑 − 𝒑)
𝒏→+∞ 𝒎𝒒
{ (𝒏(𝟗 − 𝟏𝟏𝒑 + 𝟑𝒑𝟐 ) − 𝟒 + 𝟔𝒑 − 𝟐𝒑𝟐 )𝒑(𝟏 − 𝒑) ⇔ 𝒁𝒏 → 𝒑(𝟑 − 𝒑)
𝒏→+∞
)
𝐥𝐢𝐦 𝑽(𝒁𝒏 = 𝐥𝐢𝐦 =𝟎
𝒏→+∞ 𝒏→+∞ 𝒏𝟐
𝒑
⇒ 𝒁𝒏 → 𝒑(𝟑 − 𝒑)
𝒏→+∞

Exercice 24 : (Taux de panne- Loi de Weibull- Loi des extrêmes- Loi exponentielle)
ÉNONCÉ
𝑺𝒐𝒊𝒕 𝑿 𝒖𝒏𝒆 𝒗. 𝒂. 𝒑𝒐𝒔𝒊𝒕𝒊𝒗𝒆 𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒇𝑿 𝒆𝒕 𝒅𝒆 𝒇. 𝒓. 𝑭𝑿 . 𝑶𝒏 𝒂𝒑𝒑𝒆𝒍𝒍𝒆 𝒕𝒂𝒖𝒙 𝒅𝒆 𝒑𝒂𝒏𝒏𝒆, 𝒍𝒂 𝒒𝒖𝒂𝒏𝒕𝒊𝒕é ∶
𝒇𝑿 (𝒙)
𝒕(𝒙) = , 𝒙 ∈ ℝ+ .
𝟏 − 𝑭𝑿 (𝒙)
𝒙
1) 𝑴𝒐𝒏𝒕𝒓𝒆𝒓 𝒒𝒖𝒆 𝑭𝑿 (𝒙) = 𝟏 − 𝒆−𝑻(𝒙) , 𝒐ù 𝒐𝒏 𝒂 𝒑𝒐𝒔é 𝑻(𝒙) = ∫𝟎 𝒕(𝒖)𝒅𝒖 , 𝒙 ∈ ℝ+ .
2) 𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒕(𝒙) 𝒅𝒂𝒏𝒔 𝒍𝒆𝒔 𝒄𝒂𝒔 𝒔𝒖𝒊𝒗𝒂𝒏𝒕𝒔 ∶
a) 𝑿 ↝ 𝓔(𝜽) , 𝜽 > 𝟎
𝜶−𝟏 −(𝜽𝒙𝜶 )
b) 𝑿 𝒔𝒖𝒊𝒕 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒅𝒆 𝑾𝒆𝒊𝒃𝒖𝒍𝒍 , 𝓦(𝜶, 𝜽)𝒅𝒆 𝒅. 𝒅. 𝒑. ∶ 𝒇𝑿 (𝒙) = { 𝜶𝜽𝒙 𝒆 , 𝒔𝒊 𝒙 > 𝟎 ;
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
𝜶 𝒆𝒕 𝜽 é𝒕𝒂𝒏𝒕 𝒅𝒆𝒖𝒙 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆𝒔 𝒔𝒕𝒓𝒊𝒄𝒕𝒆𝒎𝒆𝒏𝒕 𝒑𝒐𝒔𝒊𝒕𝒊𝒇.
c) 𝑿 𝒔𝒖𝒊𝒕 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒅𝒊𝒕𝒆 "𝒅𝒆𝒔 𝒆𝒙𝒕𝒓ê𝒎𝒆𝒔" 𝒅𝒆 𝒅. 𝒅. 𝒑. ∶
[𝒙−𝜽(𝒆𝒙 −𝟏)]
𝒇𝑿 (𝒙) ={ 𝜽𝒆 , 𝒔𝒊 𝒙 > 𝟎 ; 𝜽 > 𝟎 .
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
(𝐍. 𝐁. ∶ 𝒐𝒏 𝒑𝒐𝒖𝒓𝒓𝒂 𝒇𝒂𝒊𝒓𝒆 𝒍𝒆 𝒄𝒉𝒂𝒏𝒈𝒆𝒎𝒆𝒏𝒕 𝒅𝒆 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 ∶ 𝒖 = 𝜽(𝒆𝒕 − 𝟏) , 𝒑𝒐𝒖𝒓
𝒄𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝑭𝑿 (𝒙).

Corrigé
′
𝒙 𝒙 𝒇𝑿 (𝒖) 𝒙 (𝟏−𝑭𝑿 (𝒖))
1) 𝑻(𝒙) = ∫𝟎 𝒕(𝒖)𝒅𝒖 = ∫𝟎 𝒅𝒖 = − ∫𝟎 𝒅𝒖 = −[𝐥𝐧|𝟏 − 𝑭𝑿 (𝒖)|]𝒙𝟎
𝟏−𝑭𝑿 (𝒖) 𝟏−𝑭𝑿 (𝒖)

𝟎 , 𝒔𝒊 𝒙 < 𝟎
𝒙
𝑻(𝒙) = 𝐥𝐧(𝟏 − 𝑭𝑿 (𝟎)) − 𝐥𝐧(𝟏 − 𝑭𝑿 (𝒙)) , 𝑶𝒓 𝑭𝑿 (𝒙) = { ⇒ 𝑭𝑿 (𝟎) = 𝟎
∫ 𝒇𝑿 (𝒖)𝒅𝒖 , 𝒔𝒊 𝒙 ≥ 𝟎
𝟎

𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆, 𝑻(𝒙) = − 𝐥𝐧(𝟏 − 𝑭𝑿 (𝒙)) ⟺ 𝟏 − 𝑭𝑿 (𝒙) = 𝒆−𝑻(𝒙) , 𝒅′ 𝒐ù , ∀ 𝒙 ∈ ℝ+ , 𝑭𝑿 (𝒙) = 𝟏 − 𝒆−𝑻(𝒙)

2)
𝜽𝒆−𝜽𝒙 , 𝒔𝒊 𝒙 ≥ 𝟎 𝟎 , 𝒔𝒊 𝒙 < 𝟎
a) 𝑿 ↝ 𝓔(𝜽) ⟺ 𝒇𝑿 (𝒙) = { ⟺ 𝑭𝑿 (𝒙) = { 𝒙
𝟎 , 𝒔𝒊 𝒏𝒐𝒏 ∫𝟎 𝒇𝑿 (𝒖)𝒅𝒖 , 𝒔𝒊 𝒙 ≥ 𝟎
𝒙
∀ 𝒙 ∈ ℝ+ , 𝑭𝑿 (𝒙) = − ∫ (−𝜽𝒖)′ 𝒆−𝜽𝒖 𝒅𝒖 = −[𝒆−𝜽𝒖 ]𝒙𝟎 = 𝒆𝟎 − 𝒆−𝜽𝒙 = 𝟏 − 𝒆−𝜽𝒙
𝟎

𝒇𝑿 (𝒙) 𝜽𝒆−𝜽𝒙
𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆, ∀ 𝒙 ∈ ℝ+ , 𝒕(𝒙) = = −𝜽𝒙 , 𝒅𝒐ù 𝒔𝒊 𝑿 ↝ 𝓔(𝜽) , 𝒂𝒍𝒐𝒓𝒔 𝒕(𝒙) = 𝜽
𝟏 − 𝑭𝑿 (𝒙) 𝒆

353 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
b) 𝒇𝑿 (𝒙) = {𝜶𝜽𝒙 𝒆
𝜶−𝟏 −(𝜽𝒙𝜶 )
, 𝒔𝒊 𝒙 > 𝟎 ⟺ 𝑭 (𝒙) = {𝟎𝒙, 𝒔𝒊 𝒙 ≤ 𝟎
𝑿
𝟎 , 𝒔𝒊 𝒏𝒐𝒏 ∫𝟎 𝒇𝑿 (𝒖)𝒅𝒖 , 𝒔𝒊 𝒙 > 𝟎
𝒙 𝒙
𝜶−𝟏 −(𝜽𝒖𝜶 ) 𝜶 𝜶 𝒙 𝜶)
∀𝒙∈ ℝ∗+ , 𝑭𝑿 (𝒙) = ∫ 𝜶𝜽𝒖 𝒆 𝒅𝒖 = − ∫ (𝜽𝒖𝜶 )′ 𝒆−(𝜽𝒖 ) 𝒅𝒖 = −[𝒆−(𝜽𝒖 ) ]𝟎 = 𝒆𝟎 − 𝒆−(𝜽𝒙
𝟎 𝟎
𝜶)
−(𝜽𝒙𝜶 )
𝒇𝑿 (𝒙) 𝜶𝜽𝒙𝜶−𝟏 𝒆−(𝜽𝒙
∀𝒙 ∈ ℝ∗+ , 𝑭𝑿 (𝒙) =𝟏−𝒆 . 𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆, ∀ 𝒙 ∈ ℝ+ , 𝒕(𝒙) = = 𝜶 = 𝜶𝜽𝒙𝜶−𝟏
𝟏 − 𝑭𝑿 (𝒙) 𝒆−(𝜽𝒙 )

𝑫𝒐ù 𝒔𝒊 𝑿 ↝ 𝓦(𝜶, 𝜽) , 𝒂𝒍𝒐𝒓𝒔 𝒕(𝒙) = 𝜶𝜽𝒙𝜶−𝟏

c) 𝒇𝑿 (𝒙) = { 𝜽𝒆
[𝒙−𝜽(𝒆𝒙 −𝟏)]
, 𝒔𝒊 𝒙 > 𝟎 ⟺ 𝑭 (𝒙) = {𝟎𝒙, 𝒔𝒊 𝒙 ≤ 𝟎
𝑿
𝟎 , 𝒔𝒊 𝒏𝒐𝒏 ∫𝟎 𝒇𝑿 (𝒖)𝒅𝒖 , 𝒔𝒊 𝒙 > 𝟎
𝒙
𝒖 −𝟏)]
∀𝒙∈ ℝ∗+ , 𝑭𝑿 (𝒙) = ∫ 𝜽𝒆[𝒖−𝜽(𝒆 𝒅𝒖
𝟎

𝒛+𝜽 𝒛+𝜽
𝑷𝒐𝒔𝒐𝒏𝒔 𝒛 = 𝜽(𝒆𝒖 − 𝟏) ⇒ 𝒛 + 𝜽 = 𝜽𝒆𝒖 ⇒ 𝒆𝒖 = ⇒ 𝒖 = 𝐥𝐧 ( )
𝜽 𝜽
𝒅𝒛 𝒅𝒛 𝒖=𝟎⇒𝒛=𝟎
𝒆𝒕 𝒅𝒛 = [𝜽(𝒆𝒖 − 𝟏)]′ 𝒅𝒕 = 𝜽𝒆𝒖 𝒅𝒖 ⇒ 𝒅𝒖 = = ;{
𝜽𝒆 𝒖 𝒛+𝜽 𝒖 = 𝒙 ⇒ 𝒛 = 𝜽(𝒆𝒙 − 𝟏)
𝒛+𝜽
𝜽(𝒆𝒙 −𝟏) 𝒙 −𝟏)
𝒙
[𝒖−𝜽(𝒆𝒖 −𝟏)] 𝒛+𝜽 𝒅𝒛 𝜽(𝒆
𝜽𝒆𝐥𝐧( 𝜽 ) 𝒆−𝒛
𝑭𝑿 (𝒙) = ∫ 𝜽𝒆 𝒅𝒖 = ∫ 𝜽 𝐞𝐱𝐩 [𝐥𝐧 ( ) − 𝒛] =∫ 𝒅𝒛
𝟎 𝟎 𝜽 𝒛+𝜽 𝟎 𝒛+𝜽

𝜽(𝒆𝒙 −𝟏) 𝜽 (
𝒛 + 𝜽 −𝒛 𝜽(𝒆𝒙 −𝟏)
=∫ 𝜽 ) 𝒆 𝒅𝒛 = ∫ 𝒆−𝒛 𝒅𝒛 = −[𝒆−𝒛 ]𝜽(𝒆
𝒙 −𝟏) 𝒙
. 𝑨𝒊𝒏𝒔𝒊 , 𝑭𝑿 (𝒙) = 𝟏 − 𝒆−𝜽(𝒆 −𝟏)
𝟎
𝟎 𝒛+𝜽 𝟎
𝒙
𝒇𝑿 (𝒙) 𝜽𝒆[𝒙−𝜽(𝒆 −𝟏)]
𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆, ∀ 𝒙 ∈ ℝ+ , 𝒕(𝒙) = = 𝒙 = 𝜽𝒆𝒙
𝟏 − 𝑭𝑿 (𝒙) 𝒆−𝜽(𝒆 −𝟏)
[𝒙−𝜽(𝒆𝒙 −𝟏)]
𝑫𝒐ù 𝒔𝒊 𝒇𝑿 (𝒙) = { 𝜽𝒆 , 𝒔𝒊 𝒙 > 𝟎 , 𝒂𝒍𝒐𝒓𝒔 𝒕(𝒙) = 𝜽𝒆𝒙
𝟎 , 𝒔𝒊 𝒏𝒐𝒏

Exercice 25 : (Variable tronquée)

ÉNONCÉ
𝑰𝒍 𝒑𝒆𝒖𝒕 𝒂𝒓𝒓𝒊𝒗𝒆𝒓 𝒂𝒖 𝒄𝒐𝒖𝒓𝒔 𝒅’𝒖𝒏𝒆 𝒑𝒉𝒂𝒔𝒆 𝒅𝒆 𝒓𝒆𝒄𝒖𝒆𝒊𝒍 𝒅’𝒊𝒏𝒇𝒐𝒓𝒎𝒂𝒕𝒊𝒐𝒏𝒔 𝒒𝒖𝒆 𝒍𝒂 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒐𝒃𝒋𝒆𝒕 𝒅𝒆
𝒍’é𝒕𝒖𝒅𝒆 𝒏𝒆 𝒔𝒐𝒊𝒕 𝒐𝒃𝒔𝒆𝒓𝒗é𝒆 𝒒𝒖𝒆 𝒔𝒖𝒓 𝒖𝒏 𝒅𝒐𝒎𝒂𝒊𝒏𝒆 𝒑𝒍𝒖𝒔 𝒓𝒆𝒔𝒕𝒓𝒆𝒊𝒏𝒕 𝒒𝒖𝒆 𝒄𝒆𝒍𝒖𝒊 𝒔𝒖𝒓 𝒍𝒆𝒒𝒖𝒆𝒍 𝒆𝒍𝒍𝒆 𝒅𝒐𝒊𝒕
𝒑𝒓𝒆𝒏𝒅𝒓𝒆, 𝒕𝒉é𝒐𝒓𝒊𝒒𝒖𝒆𝒎𝒆𝒏𝒕, 𝒔𝒆𝒔 𝒗𝒂𝒍𝒆𝒖𝒓𝒔 : 𝒄. ‑à‑𝒅. 𝒒𝒖’𝒂𝒖‑𝒅𝒆𝒍à 𝒅’𝒖𝒏 𝒄𝒆𝒓𝒕𝒂𝒊𝒏 𝒔𝒆𝒖𝒊𝒍 𝜽 , 𝒐𝒏 𝒏′ 𝒐𝒃𝒔𝒆𝒓𝒗𝒆
𝒂𝒖𝒄𝒖𝒏𝒆 𝒅𝒐𝒏𝒏é𝒆. 𝑶𝒏 𝒔′ 𝒊𝒏𝒕é𝒓𝒆𝒔𝒔𝒆 à 𝒄𝒆 𝒕𝒚𝒑𝒆 𝒅𝒆 𝒔𝒊𝒕𝒖𝒂𝒕𝒊𝒐𝒏. 𝑺𝒐𝒊𝒕 𝑿 𝒖𝒏𝒆 𝒗. 𝒂. 𝒓é𝒆𝒍𝒍𝒆 𝒅𝒆 𝒇. 𝒓. 𝑭𝑿 (𝒙)
𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒆𝒕 𝒔𝒕𝒓𝒊𝒄𝒕𝒆𝒎𝒆𝒏𝒕 𝒄𝒓𝒐𝒊𝒔𝒔𝒂𝒏𝒕𝒆, 𝒆𝒕 𝒅𝒆 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒇𝑿 (𝒙) , 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 .
1) 𝑺𝒐𝒊𝒕 𝒀𝜽 𝒍𝒂 𝒗. 𝒂. 𝑿 𝒐𝒃𝒔𝒆𝒓𝒗é𝒆 𝒔𝒖𝒓 ]−∞, 𝜽].
𝑪𝒂𝒍𝒄𝒖𝒍𝒆𝒓 𝒍𝒂 𝒇. 𝒓. 𝑮𝜽 𝒅𝒆 𝒀𝜽 . 𝑬𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒔𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒈𝜽 .

354 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
2) 𝑺𝒐𝒊𝒆𝒏𝒕 𝒀𝟏 , 𝒀𝟐 , … , 𝒀𝒏 𝒏 𝒗. 𝒂. 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 𝒔𝒖𝒊𝒗𝒂𝒏𝒕 𝒍𝒂 𝒎ê𝒎𝒆 𝒍𝒐𝒊 𝒒𝒖𝒆 𝒀𝜽 . 𝑸𝒖𝒆𝒍𝒍𝒆 𝒆𝒔𝒕
𝒍𝒂 𝒅𝒆𝒏𝒔𝒊𝒕é 𝒈𝒏,𝜽 𝒅𝒆 (𝒀𝟏 , 𝒀𝟐 , … , 𝒀𝒏 ) ?
3) 𝑻𝒓𝒂𝒄𝒆𝒓 𝒍𝒂 𝒄𝒐𝒖𝒓𝒃𝒆 𝒓𝒆𝒑𝒓é𝒔𝒆𝒏𝒕𝒂𝒕𝒊𝒗𝒆 𝒅𝒆 𝒈𝒏,𝜽 (𝒚𝟏 , 𝒚𝟐 , … , 𝒚𝒏 ) 𝒆𝒏 𝒇𝒐𝒏𝒄𝒕𝒊𝒐𝒏 𝒅𝒆 𝜽 . 𝑬𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆
𝒒𝒖′ 𝒊𝒍 𝒆𝒙𝒊𝒔𝒕𝒆 𝒖𝒏 𝒎𝒂𝒙𝒊𝒎𝒖𝒎 𝒖𝒏𝒊𝒒𝒖𝒆 𝒂𝒕𝒕𝒆𝒊𝒏𝒕 𝒆𝒏 𝒖𝒏 𝒑𝒐𝒊𝒏𝒕 𝒔𝒏 (𝒚𝟏 , 𝒚𝟐 , … , 𝒚𝒏 ) 𝒒𝒖𝒆 𝒍′ 𝒐𝒏
𝒑𝒓é𝒄𝒊𝒔𝒆𝒓𝒂 .
4) 𝑺𝒐𝒊𝒕 𝑺𝒏 𝒍𝒂 𝒗. 𝒂. 𝒂𝒔𝒔𝒐𝒄𝒊é𝒆 à 𝒔𝒏 . 𝑸𝒖𝒆𝒍𝒍𝒆 𝒆𝒔𝒕 𝒍𝒂 𝒍𝒐𝒊 𝒅𝒆 𝑺𝒏 ?
5) 𝑴𝒐𝒏𝒕𝒓𝒆𝒓 𝒒𝒖𝒆 𝑺𝒏 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕é 𝒗𝒆𝒓𝒔 𝜽
6) 𝑺𝒐𝒊𝒕 𝒉 = 𝐥𝐧(𝑭) 𝒆𝒕 𝑼𝒏 = 𝒏(𝜽 − 𝑺𝒏 ) .
a) 𝑬𝒙𝒑𝒓𝒊𝒎𝒆𝒓 𝑷(𝑼𝒏 < 𝒖) , 𝒖 ∈ ℝ
b) 𝑬𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒍𝒂 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒄𝒆 𝒆𝒏 𝒍𝒐𝒊 𝒅𝒆 𝑼𝒏 , 𝒒𝒖𝒂𝒏𝒅 𝒏 𝒕𝒆𝒏𝒅 𝒗𝒆𝒓𝒔 𝒍′ 𝒊𝒏𝒇𝒊𝒏𝒊, 𝒗𝒆𝒓𝒔 𝒖𝒏𝒆
𝒍𝒐𝒊 𝒒𝒖𝒆 𝒍′ 𝒐𝒏 𝒑𝒓é𝒄𝒊𝒔𝒆𝒓𝒂 (𝒐𝒏 𝒂𝒅𝒎𝒆𝒕𝒕𝒓𝒂 𝒒𝒖𝒆 𝒉′ (𝜽) ≠ 𝟎)

Corrigé
1) 𝑮𝜽 (𝒚) = 𝑷(𝒀𝜽 ≤ 𝒚) = 𝑷((𝑿 ≤ 𝒚)|(𝑿 ∈ ]−∞, 𝜽])) = 𝑷((𝑿 ∈ ]−∞, 𝒚])|(𝑿 ∈ ]−∞, 𝜽]))
𝑷[𝑿 ∈ (]−∞, 𝒚] ∩ ]−∞, 𝜽])]
=
𝑷(𝑿 ≤ 𝜽)
𝑷(𝑿 ≤ 𝜽)
∙ 𝑺𝒊 𝒚 ≥ 𝜽 ⇒ ]−∞, 𝒚] ∩ ]−∞, 𝜽] = ]−∞, 𝜽] 𝒆𝒕 𝑮𝜽 (𝒚) = =𝟏
𝑷(𝑿 ≤ 𝜽)
𝑷(𝑿 ≤ 𝒚) 𝑭𝑿 (𝒚)
∙ 𝑺𝒊 𝒚 < 𝜽 ⇒ ]−∞, 𝒚] ∩ ]−∞, 𝜽] = ]−∞, 𝒚] 𝒆𝒕 𝑮𝜽 (𝒚) = =
𝑷(𝑿 ≤ 𝜽) 𝑭𝑿 (𝜽)

𝟏 , 𝒔𝒊 𝒚 ≥ 𝜽 𝒇𝑿 (𝒚)
𝒅𝑮𝜽 (𝒚) , 𝒔𝒊 𝒚 < 𝜽
′ 𝑭
𝑫 𝒐ù 𝑮𝜽 (𝒚) = { 𝑿 (𝒚) . 𝑶𝒓 𝒈𝜽 (𝒚) = ⇒ 𝒈𝜽 (𝒚) = {𝑭𝑿 (𝜽)
, 𝒔𝒊 𝒚 < 𝜽 𝒅𝒚
𝑭𝑿 (𝜽) 𝟎, 𝒔𝒊 𝒏𝒐𝒏

2) (𝒀𝟏 , 𝒀𝟐 , … , 𝒀𝒏 ) 𝒆𝒔𝒕 𝒖𝒏𝒆 𝒔𝒖𝒊𝒕𝒆 𝒅𝒆 𝒗. 𝒂. 𝒊. 𝒊. 𝒅. , 𝒑𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 ∶

𝒏
𝒏 𝒇𝑿 (𝒚𝒊 )
∏ , 𝒔𝒊 ∀𝒊 = 𝟏, 𝟐, … , 𝒏, 𝒚𝒊 < 𝜽
𝒈𝒏,𝜽 (𝒚𝟏 , 𝒚𝟐 , … , 𝒚𝒏 ) = ∏ 𝒈𝜽 (𝒚𝒊 ) = { 𝑭𝑿 (𝜽)
𝒊=𝟏
𝒊=𝟏
𝟎, 𝒔𝒊 𝒏𝒐𝒏
𝒏
𝟏
∏ 𝒇𝑿 (𝒚𝒊 ) , 𝒔𝒊 ( 𝐬𝐮𝐩 𝒚𝒊 ) < 𝜽
𝑫′ 𝒐ù 𝒈𝒏,𝜽 (𝒚𝟏 , 𝒚𝟐 , … , 𝒚𝒏 ) = {(𝑭 (𝜽))𝒏 𝟏≤𝒊≤𝒏
𝑿 𝒊=𝟏
𝟎, 𝒔𝒊 𝒏𝒐𝒏
′
𝒅𝒈𝒏,𝜽 (𝜽) (𝑭𝑿 (𝜽))
3) = −𝒏(∏𝒏𝒊=𝟏 𝒇𝑿 (𝒚𝒊 )) 𝒏+𝟏 ≤ 𝟎 , 𝒄𝒂𝒓▪∀𝒊 ; 𝒇𝑿 (𝒚𝒊 ) ≥ 𝟎 ⇒ ∏𝒏𝒊=𝟏 𝒇𝑿 (𝒚𝒊 ) ≥ 𝟎
𝒅𝜽 (𝑭𝑿 (𝜽))

′
▪𝟎 ≤ 𝑭𝑿 (𝜽) ≤ 𝟏 𝒆𝒕 ▪𝑭𝑿 (𝜽) 𝒆𝒔𝒕 𝒔𝒕𝒓𝒊𝒄𝒕𝒆𝒎𝒆𝒏𝒕 𝒄𝒓𝒐𝒊𝒔𝒔𝒂𝒏𝒕𝒆 , 𝒅𝒐𝒏𝒄 (𝑭𝑿 (𝜽)) > 𝟎
𝒏
𝑨𝒊𝒏𝒔𝒊 𝒈𝒏,𝜽 (𝜽) 𝒆𝒔𝒕 𝒅é𝒄𝒓𝒐𝒊𝒔𝒔𝒂𝒏𝒕𝒆 . 𝒐𝒏 𝒂 𝒂𝒖𝒔𝒔𝒊 ∶ 𝐥𝐢𝐦 (𝑭𝑿 (𝜽)) = 𝟏
𝜽→+∞

355 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014

𝑫’𝒐ù 𝒈𝒏,𝜽 (𝒚𝟏 , 𝒚𝟐 , … , 𝒚𝒏 ) 𝒂𝒕𝒕𝒆𝒊𝒏𝒕 𝒔𝒐𝒏 𝒎𝒂𝒙𝒊𝒎𝒖𝒎 𝒂𝒖 𝒑𝒐𝒊𝒏𝒕 𝒅’𝒂𝒃𝒔𝒄𝒊𝒔𝒔𝒆 𝒔𝒏 (𝒚𝟏 , 𝒚𝟐 , … , 𝒚𝒏 ) = 𝐬𝐮𝐩 𝒚𝒊
𝟏≤𝒊≤𝒏

4) 𝑭𝑺𝒏 (𝒔) = 𝑷(𝑺𝒏 ≤ 𝒔) = 𝑷 ( 𝐬𝐮𝐩 𝒀𝒊 ≤ 𝒔) = 𝑷(𝒀𝟏 ≤ 𝒔, 𝒀𝟐 ≤ 𝒔 , … , 𝒀𝒏 ≤ 𝒔) = ∏𝒏𝒊=𝟏 𝑷(𝒀𝒊 ≤ 𝒔)

𝟏≤𝒊≤𝒏
𝒏 𝒏 𝒏
𝑭𝑿 (𝒔) 𝑭𝑿 (𝒔)
= ∏ 𝑮𝜽 (𝒔) = ∏ =( )
𝑭𝑿 (𝜽) 𝑭𝑿 (𝜽)
𝒊=𝟏 𝒊=𝟏

𝟏 , 𝒔𝒊 𝒔 ≥ 𝜽
𝒏
𝑭𝑺𝒏 (𝒔) = { 𝑭𝑿 (𝒔)
( ) , 𝒔𝒊 𝒚 < 𝜽
𝑭𝑿 (𝜽)
𝒅𝑭𝑺𝒏 (𝒔) 𝒏𝒇𝑿 (𝒔) 𝒏−𝟏
𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆, 𝒇𝑺𝒏 (𝒔) = = 𝒏 (𝑭𝑿 (𝒔))
𝒅𝒔 (𝑭𝑿 (𝜽))

𝒏𝒇𝑿 (𝒔) 𝒏−𝟏

𝒏 (𝑭𝑿 (𝒔)) , 𝒔𝒊 𝒔 < 𝜽
⇒ 𝒇𝑺𝒏 (𝒔) = {(𝑭𝑿 (𝜽))
𝟎, 𝒔𝒊 𝒏𝒐𝒏

5) 𝒀𝜽 é𝒕𝒂𝒏𝒕 𝒍𝒂 𝒗. 𝒂. 𝑿 𝒐𝒃𝒔𝒆𝒓𝒗é𝒆 𝒔𝒖𝒓 ]−∞, 𝜽], 𝑺𝒏 = 𝐬𝐮𝐩 𝒀𝒊 ⇒ 𝑷(𝑺𝒏 < 𝜽) = 𝟏

𝟏≤𝒊≤𝒏

𝑶𝒓 ∀𝝐 > 𝟎 , 𝐥𝐢𝐦 𝑷(|𝑺𝒏 − 𝜽| ≥ 𝝐) = 𝐥𝐢𝐦 [𝑷 (𝑺 𝐥𝐢𝐦 𝑭𝑺𝒏 (𝜽 − 𝝐)

⏟ 𝒏 ≥ 𝜽 + 𝝐) + 𝑷(𝑺𝒏 ≤ 𝜽 − 𝝐)] = 𝒏→+∞
𝒏→+∞ 𝒏→+∞
∅
𝒏
𝑭𝑿 (𝜽 − 𝝐) 𝑭𝑿 (𝜽 − 𝝐)
∀𝝐 > 𝟎 , 𝐥𝐢𝐦 𝑷(|𝑺𝒏 − 𝜽| ≥ 𝝐) = 𝐥𝐢𝐦 ( ) = 𝟎 , 𝒄𝒂𝒓 𝟎 ≤ ≤𝟏
𝒏→+∞ 𝒏→+∞ 𝑭𝑿 (𝜽) 𝑭𝑿 (𝜽)
𝒑
𝑫′ 𝒐ù 𝑺𝒏 → 𝜽
𝒏→+∞

6)
𝑼 = 𝒏(𝜽 − 𝑺𝒏 )
a) { 𝒏 ⇒ 𝑼𝒏 (𝛀) = ℝ∗+ ,
𝑺𝒏 (𝛀) = ]−∞, 𝜽[
𝒖 𝒖 𝒖
𝑷(𝑼𝒏 < 𝒖) = 𝑷(𝒏(𝜽 − 𝑺𝒏 ) < 𝒖) = 𝑷 (𝜽 − 𝑺𝒏 < ) = 𝑷 (𝜽 − < 𝑺𝒏 ) = 𝟏 − 𝑷 (𝑺𝒏 ≤ 𝜽 − )
𝒏 𝒏 𝒏

356 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒖 𝒏
𝒖 𝑭𝑿 (𝜽 − 𝒏)
𝑷(𝑼𝒏 < 𝒖) = 𝟏 − 𝑭𝑺𝒏 (𝜽 − ) . 𝑫′ 𝒐ù, ∀𝒖 > 𝟎, 𝑷(𝑼𝒏 < 𝒖) = 𝟏 − ( )
𝒏 𝑭𝑿 (𝜽)

𝟏 , 𝒔𝒊 𝒖 ≤ 𝟎
𝒖 𝒏
b) 𝑷(𝑼𝒏 < 𝒖) = 𝑭𝑼𝒏 (𝒖) = { 𝑭𝑿 (𝜽− )
𝒏
𝟏−( ) , 𝒔𝒊 𝒖 > 𝟎
𝑭𝑿 (𝜽)

𝒏
𝒖 𝒖
𝒖 𝒏 𝐥𝐧(
𝑭𝑿 (𝜽− )
𝒏 ) 𝒏 𝐥𝐧(
𝑭𝑿 (𝜽− )
𝒏 )
𝑭𝑿 (𝜽 − 𝒏) 𝑭𝑿 (𝜽) 𝑭𝑿 (𝜽)
𝐥𝐢𝐦 𝑭𝑼𝒏 (𝒖) = 𝐥𝐢𝐦 [𝟏 − ( ) ] = 𝐥𝐢𝐦 𝟏 − 𝒆 = 𝐥𝐢𝐦 𝟏 − 𝒆
𝒏→+∞ 𝒏→+∞ 𝑭𝑿 (𝜽) 𝒏→+∞ 𝒏→+∞

[ ] [ ]
𝒖
𝒏[𝐥𝐧(𝑭𝑿 (𝜽− ))−𝐥𝐧(𝑭𝑿 (𝜽))]
= 𝐥𝐢𝐦 [𝟏 − 𝒆 𝒏 ] ; 𝒐𝒓 𝒉 = 𝐥𝐧(𝑭) , 𝒑𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆 ∶
𝒏→+∞

𝒖 𝒖
𝒉(𝜽− )−𝒉(𝜽) 𝒉(𝜽− )−𝒉(𝜽)
[ 𝒏 ] 𝒖[ 𝒏 ]
𝒖 𝟏⁄𝒏 𝒖⁄𝒏
𝐥𝐢𝐦 𝑭𝑼𝒏 (𝒖) = 𝐥𝐢𝐦 [𝟏 − 𝒆𝒏[𝒉(𝜽−𝒏)−𝒉(𝜽)] ] = 𝐥𝐢𝐦 𝟏 − 𝒆 = 𝐥𝐢𝐦 𝟏 − 𝒆
𝒏→+∞ 𝒏→+∞ 𝒏→+∞ 𝒏→+∞
[ ] [ ]
𝒉(𝜽−𝒕)−𝒉(𝜽) 𝒖
𝒖[ ]
= 𝐥𝐢𝐦 [𝟏 − 𝒆 𝒕 ] , 𝒂𝒗𝒆𝒄 ,𝒕 = 𝒆𝒕 (𝒔𝒊, 𝒏 → +∞, 𝒂𝒍𝒐𝒓𝒔, 𝒕 → 𝟎)
𝒕→𝟎 𝒏
𝒉(𝒗)−𝒉(𝜽)
−𝒖[ ]
= 𝐥𝐢𝐦 [𝟏 − 𝒆 𝒗−𝜽 ] , 𝒂𝒗𝒆𝒄 , 𝒗 = 𝜽 − 𝒕 𝒆𝒕 (𝒔𝒊, 𝒕 → 𝟎, 𝒂𝒍𝒐𝒓𝒔, 𝒗 → 𝜽)
𝒗→𝜽

′
𝐥𝐢𝐦 𝑭𝑼𝒏 (𝒖) = 𝟏 − 𝒆−[𝒉 (𝜽)]𝒖
𝒏→+∞

′ −[𝒉 (𝜽)]𝒖 ′
𝑺𝒐𝒊𝒕 𝒍𝒂 𝒗. 𝒂. 𝑼 ↝ 𝓔(𝒉′ (𝜽)) ⟺ 𝒇𝑼 (𝒖) = {𝒉 (𝜽)𝒆 , 𝒔𝒊 𝒖 ∈ [𝟎, +∞[
𝟎 , 𝒔𝒊 𝒏𝒐𝒏
𝟎 , 𝒔𝒊 𝒖 < 𝟎
⟺ 𝑭𝑼 (𝒖) = { ′
𝟏 − 𝒆−[𝒉 (𝜽)]𝒖 , 𝒔𝒊 𝒖 ≥ 𝟎
𝓛
𝑫′ 𝒐ù 𝑼𝒏 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆 𝒆𝒏 𝒍𝒐𝒊 𝒗𝒆𝒓𝒔 𝑼 ∶ 𝑼𝒏 → 𝑼 , 𝒐ù 𝑼 ↝ 𝓔(𝒉′ (𝜽))
𝒏→+∞

Exercice 26 : (Théorème de Fisher-Cochran)

ÉNONCÉ
𝑳𝒆 𝒃𝒖𝒕 𝒅𝒆 𝒄𝒆𝒕 𝒆𝒙𝒆𝒓𝒄𝒊𝒄𝒆 𝒆𝒔𝒕 𝒅𝒆 𝒅é𝒎𝒐𝒏𝒕𝒓𝒆𝒓 𝒅’𝒖𝒏𝒆 𝒂𝒖𝒕𝒓𝒆 𝒇𝒂ç𝒐𝒏 𝒍𝒆 𝒕𝒉é𝒐𝒓è𝒎𝒆 𝒅𝒆 𝑭𝒊𝒔𝒉𝒆𝒓
𝒏
𝟏
̅ 𝒏 𝒆𝒕
é𝒕𝒂𝒃𝒍𝒊𝒔𝒔𝒂𝒏𝒕 𝒍’𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒄𝒆 𝒅𝒆 𝑿 𝑺𝟐𝒏 = ̅ 𝒏 )𝟐 .
∑(𝑿𝒊 − 𝑿
𝒏−𝟏
𝒊=𝟏

𝑺𝒐𝒊𝒕 (𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝒏 ) 𝒖𝒏 é𝒄𝒉𝒂𝒏𝒕𝒊𝒍𝒍𝒐𝒏 𝒊. 𝒊. 𝒅 𝒊𝒔𝒔𝒖 𝒅𝒆 𝒍𝒂 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆 𝓝(𝝁, 𝝈𝟐 ), 𝒐𝒏 𝒔𝒂𝒊𝒕 𝒒𝒖𝒆 𝒍𝒂

357 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝝈𝟐
̅ 𝒏 ↝ 𝓝 (𝝁,
𝒎𝒐𝒚𝒆𝒏𝒏𝒆 𝒆𝒎𝒑𝒊𝒓𝒊𝒒𝒖𝒆 𝑿 )
𝒏

1) 𝑶𝒏 𝒂𝒅𝒎𝒆𝒕 𝒒𝒖𝒆, ∀𝒊, (𝑿𝒊 − 𝑿 ̅ 𝒏 ) 𝒔𝒖𝒊𝒕 𝒖𝒏𝒆 𝒍𝒐𝒊 𝒏𝒐𝒓𝒎𝒂𝒍𝒆. 𝑫é𝒕𝒆𝒓𝒎𝒊𝒏𝒆𝒓 𝒍𝒆𝒔 𝒑𝒂𝒓𝒂𝒎è𝒕𝒓𝒆𝒔 𝒅𝒆
𝒄𝒆𝒕𝒕𝒆 𝒍𝒐𝒊
2) 𝑴𝒐𝒏𝒕𝒓𝒆𝒓 𝒒𝒖𝒆 𝑪𝒐𝒗(𝑿 ̅ 𝒏 , (𝑿𝒊 − 𝑿 ̅ 𝒏 )) = 𝟎 𝒆𝒕 𝒆𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒒𝒖𝒆 𝒍𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝒂𝒍é𝒂𝒕𝒐𝒊𝒓𝒆𝒔 𝑿
̅ 𝒏 𝒆𝒕
̅ 𝒏 ) 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔
(𝑿𝒊 − 𝑿
3) 𝑭𝒊𝒏𝒂𝒍𝒆𝒎𝒆𝒏𝒕 𝒆𝒏 𝒅é𝒅𝒖𝒊𝒓𝒆 𝒒𝒖𝒆 𝑿 ̅ 𝒏 𝒆𝒕 𝑺𝟐𝒏 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔

Corrigé
1)
̅ 𝒏 ) = 𝑬(𝑿𝒊 ) − 𝑬(𝑿
∙ 𝑬(𝑿𝒊 − 𝑿 ̅ 𝒏) = 𝝁 − 𝝁 = 𝟎
𝒏 𝒏 𝒏
𝟏 𝟏 𝟏 𝒏−𝟏 𝟏
̅ 𝒏 = 𝑿𝒊 − ∑ 𝑿𝒌 = (𝑿𝒊 − 𝑿𝒊 ) − ∑ 𝑿𝒌 = (
∙ 𝑿𝒊 − 𝑿 ) 𝑿𝒊 − ∑ 𝑿𝒌
𝒏 𝒏 𝒏 𝒏 𝒏
𝒌=𝟏 𝒌≠𝒊 𝒌≠𝒊

𝒏
𝒏−𝟏 𝟏
̅ 𝒏 ) = 𝑽 ((
∙ 𝑽(𝑿𝒊 − 𝑿 ) 𝑿𝒊 − ∑ 𝑿𝒌 )
𝒏 𝒏
𝒌≠𝒊

𝒏 𝒏
𝒏−𝟏 𝟐 𝟏 𝟐(𝒏 − 𝟏)
=( ) 𝑽(𝑿𝒊 ) + 𝟐 𝑽 (∑ 𝑿𝒌 ) − 𝑪𝒐𝒗 (𝑿𝒊 , ∑ 𝑿𝒌 )
𝒏 𝒏 𝒏𝟐
𝒌≠𝒊 𝒌≠𝒊

▪(𝑿𝒊 )𝟏≤𝒊≤𝒏 é𝒄𝒉𝒂𝒏𝒕𝒊𝒍𝒍𝒐𝒏 𝒊. 𝒊. 𝒅. 𝒅𝒆 𝓝(𝝁, 𝝈𝟐 )

𝒏 𝒏

⇒ 𝑽 (∑ 𝑿𝒌 ) = (∑ 𝑽𝒂𝒓(𝑿𝒌 )) − 𝑽(𝑿𝒊 ) = 𝒏𝝈𝟐 − 𝝈𝟐 = (𝒏 − 𝟏)𝝈𝟐

𝒌≠𝒊 𝒌=𝟏
𝒏 𝒏
𝟐)
▪(𝑿𝒊 )𝟏≤𝒊≤𝒏 é𝒄𝒉𝒂𝒏𝒕𝒊𝒍𝒍𝒐𝒏 𝒊. 𝒊. 𝒅. 𝒅𝒆 𝓝(𝝁, 𝝈 ⇒ 𝑪𝒐𝒗 (𝑿𝒊 , ∑ 𝑿𝒌 ) = ∑ 𝑪𝒐𝒗(𝑿𝒊 , 𝑿𝒌 ) = 𝟎
𝒌≠𝒊 𝒌≠𝒊
𝒏 𝒏
𝒏−𝟏 𝟐 𝟏 𝟐(𝒏 − 𝟏)
̅ )
𝑷𝒂𝒓 𝒍𝒂 𝒔𝒖𝒊𝒕𝒆, 𝑽(𝑿𝒊 − 𝑿𝒏 = ( ) 𝑽(𝑿𝒊 ) + 𝟐 𝑽 (∑ 𝑿𝒌 ) − 𝑪𝒐𝒗 (𝑿𝒊 , ∑ 𝑿𝒌 )
𝒏 𝒏 𝒏𝟐
𝒌≠𝒊 𝒌≠𝒊

𝒏−𝟏 𝟐 𝟐 𝒏−𝟏 𝒏−𝟏 𝒏−𝟏 𝟏 𝟐 𝒏−𝟏 𝟐

=( ) 𝝈 + ( 𝟐 ) 𝝈𝟐 = ( )( + )𝝈 = ( )𝝈
𝒏 𝒏 𝒏 𝒏 𝒏 𝒏
𝒏−𝟏 𝟐
̅ 𝒏 ) ↝ 𝓝 (𝟎, (
𝑫′ 𝒐ù (𝑿𝒊 − 𝑿 )𝝈 )
𝒏
𝟐
̅ 𝒏 , (𝑿𝒊 − 𝑿
2) 𝑪𝒐𝒗(𝑿 ̅ 𝒏 )) = 𝑪𝒐𝒗(𝑿
̅ 𝒏 , 𝑿𝒊 ) − 𝑪𝒐𝒗(𝑿
̅ 𝒏, 𝑿
̅ 𝒏 ) = 𝑪𝒐𝒗(𝑿
̅ 𝒏 , 𝑿𝒊 ) − 𝑽(𝑿 ̅ 𝒏 , 𝑿𝒊 ) − 𝝈
̅ 𝒏 ) = 𝑪𝒐𝒗(𝑿
𝒏
𝒏 𝒏 𝒏
𝟏 𝟏 𝟏
̅ 𝒏 ) = 𝑪𝒐𝒗 [𝑿𝒊 , ∑ 𝑿𝒌 ] = 𝑪𝒐𝒗 [𝑿𝒊 , (𝑿𝒊 + ∑ 𝑿𝒌 )] = 𝑪𝒐𝒗 [𝑿𝒊 , (𝑿𝒊 + ∑ 𝑿𝒌 )]
▪𝑪𝒐𝒗(𝑿𝒊 , 𝑿
𝒏 𝒏 𝒏
𝒌=𝟏 𝒌≠𝒊 𝒌≠𝒊

358 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
𝒏
𝟏 𝟏 𝝈𝟐
= 𝑪𝒐𝒗(𝑿𝒊 , 𝑿𝒊 ) + 𝑪𝒐𝒗 (𝑿𝒊 , ∑ 𝑿𝒌 ) = 𝑽(𝑿𝒊 ) =
𝒏 ⏟
𝒏 𝒏
𝒌≠𝒊
[ 𝟎 ]

̅ 𝒏 , (𝑿𝒊 − 𝑿
𝑫′ 𝒐ù ∶ 𝑪𝒐𝒗(𝑿 ̅ 𝒏 )) = 𝟎 ; 𝒍𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔𝑿
̅ 𝒏 , (𝑿𝒊 − 𝑿
̅ 𝒏 ) 𝒔𝒐𝒏𝒕 𝒏𝒐𝒏‑𝒄𝒐𝒓𝒓é𝒍é𝒆𝒔

𝝈𝟐
̅ 𝒏 ↝ 𝓝 (𝝁,
𝑿 )
𝒏
▪𝑶𝒏 𝒂 ∶ ̅ 𝒏 , (𝑿𝒊 − 𝑿
𝒏 − 𝟏 𝟐 ⟺ 𝑳𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝑿 ̅ 𝒏 ) 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔
̅ 𝒏 ) ↝ 𝓝 (𝟎, (
(𝑿𝒊 − 𝑿 )𝝈 )
𝒏
̅ ̅
{𝑪𝒐𝒗(𝑿𝒏 , (𝑿𝒊 − 𝑿𝒏 )) = 𝟎
3)
̅ 𝒏 )𝟐 = 𝒇((𝑿𝒊 − 𝑿
▪𝑺𝒐𝒊𝒕 𝒇(𝒖) = 𝒖𝟐 , 𝒅𝒐𝒏𝒄 (𝑿𝒊 − 𝑿 ̅ 𝒏 )), 𝒇 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒔𝒖𝒓 ℝ (𝟏)

̅ 𝒏 = 𝒈(𝑿
▪𝒈(𝒖) = 𝒖, 𝒅𝒐𝒏𝒄 𝑿 ̅ 𝒏 ), 𝒈 𝒄𝒐𝒏𝒕𝒊𝒏𝒖𝒆 𝒔𝒖𝒓 ℝ (𝟐)

𝒏−𝟏 𝟐 𝒏 ̅𝒏
𝑿𝒊 − 𝑿 ̅ 𝒏 )𝟐
𝒏(𝑿𝒊 − 𝑿
̅ 𝒏 ) ↝ 𝓝 (𝟎, (
▪(𝑿𝒊 − 𝑿 )𝝈 ) ⇒ √ ( ) ↝ 𝓝(𝟎, 𝟏) ⇒ ↝ 𝝌𝟐 (𝟏)
𝒏 𝒏−𝟏 𝝈 (𝒏 − 𝟏)𝝈𝟐
̅ 𝒏 )𝟐
𝒏(𝑿𝒊 − 𝑿 𝒏
𝑨𝒊𝒏𝒔𝒊, 𝑬 [ ] = 𝟏 ⇒ ̅ 𝒏 )𝟐 ) = 𝟏
𝑬((𝑿𝒊 − 𝑿
(𝒏 − 𝟏)𝝈𝟐 (𝒏 − 𝟏)𝝈𝟐

(𝒏 − 𝟏)𝝈𝟐
̅ 𝒏 ))) = 𝑬((𝑿𝒊 − 𝑿
𝑪𝒆 𝒒𝒖𝒊 𝒅𝒐𝒏𝒏𝒆 ∶ 𝑬 (𝒇((𝑿𝒊 − 𝑿 ̅ 𝒏 )𝟐 ) = , 𝒆𝒙𝒊𝒔𝒕𝒆 𝒆𝒕 𝒇𝒊𝒏𝒊𝒆 (𝟑)
𝒏
𝝈𝟐
̅ 𝒏 ↝ 𝓝 (𝝁,
▪𝑿 ̅ 𝒏 )) = 𝑬(𝑿
) ⇒ 𝑬(𝒈(𝑿 ̅ 𝒏 ) = 𝝁 , 𝒆𝒙𝒊𝒔𝒕𝒆 𝒆𝒕 𝒇𝒊𝒏𝒊𝒆 (𝟒)
𝒏
̅ 𝒏 , (𝑿𝒊 − 𝑿
▪𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝑿 ̅ 𝒏 ) 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 (𝟓)
̅ 𝒏 , (𝑿𝒊 − 𝑿
(𝟏) + (𝟐) + (𝟑) + (𝟒) + (𝟓) ⇒ 𝑳𝒆𝒔 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆𝒔 𝑿 ̅ 𝒏 )𝟐 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔
𝒏
𝟏
̅ 𝒏 , (𝑿𝒊 − 𝑿
𝑳𝒆𝒔 𝒗. 𝒂. 𝑿 ̅ 𝒏 )𝟐 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔 ⇒ 𝒈(𝑿
̅ 𝒏) = 𝑿
̅ 𝒏 𝒆𝒕 𝒉(𝑿𝒊 − 𝑿
̅ 𝒏 )𝟐 = ̅ 𝒏 )𝟐
∑(𝑿𝒊 − 𝑿
𝒏−𝟏
𝒊=𝟏

𝒔𝒐𝒏𝒕 𝒂𝒖𝒔𝒔𝒊 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔.

𝒏 𝒏
′
𝟏
̅ 𝒏 = ∑ 𝑿𝒌 𝒆𝒕 𝑺𝟐𝒏 =
𝟏 ̅ 𝒏 )𝟐 𝒔𝒐𝒏𝒕 𝒊𝒏𝒅é𝒑𝒆𝒏𝒅𝒂𝒏𝒕𝒆𝒔
𝑫 𝒐ù 𝑿 ∑(𝑿𝒊 − 𝑿
𝒏 𝒏−𝟏
𝒌=𝟏 𝒊=𝟏

359 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014

Variable aléatoire continu………………………………………………………………………………….. 284

• Définition∶
• Fonction densité de probabilité∶
a. Définition 1 ∶
b. Définition 2∶
 Remarque ∶
• Fonction de répartition ∶
• Propriétés de la fonction de répartition ∶
• Les caractéristiques de tendance centrale∶
a. La médiane ∶
b. Les quartiles∶
c. Les déciles∶
d. Les centiles∶
e. Le mode∶
• Espérance mathématique ∶
• Espérance d’une fonction d’une v.a.∶
 Remarque ∶
• Propriétés de l'espérance mathématique∶
 Généralisation ∶
d. Positivité de l’espérance ∶
e. Inégalité de Jensen∶
• Variance ∶
• Propriétés de la variance∶
c. Formule de Koenig ∶
• Les moments non‐centrés d'ordre r∶
• Les moments centrés d'ordre r∶
• Relations entre moments non‑centrés et moments centrés∶
a. Moments centrés en fonction des moments non‑centrés∶
b. Moments non‑centrés en fonction des moments centrés∶
• Le coefficient d'asymétrie:
• Le coefficient d'aplatissement (kurtosis):
• Fonction génératrice des moments (ou transformée de Laplace) ∶
• Propriétés de la fonction génératrice des moments∶
• Changement de variables∶
Lois continues usuelles..……………………………………………………………………………………..289
• Loi Uniforme Continue (ou sur un intervalla [a,b]) 𝑼[𝒂,𝒃] ∶
 Fonction densité de probabilité ∶
 Fonction de répartition∶
 Espérance mathématique∶
 Variance∶
 Médiane∶
 Coefficient d'asymétrie∶
 Coefficient d'aplatissement (kurtosis normalisé) :
 Fonction génératrice des moments∶
 Remarques∶
• Loi Exponentielle E(λ) ∶
 Fonction densité de probabilité ∶
 Fonction de répartition∶
 Espérance mathématique∶
 Variance∶
 Coefficient d'asymétrie∶
 Coefficient d'aplatissement (kurtosis normalisé) :
 Fonction génératrice des moments∶
 Stabilité∶
 Remarques∶
 Utilisation∶
 Propriétés d'absence de mémoire∶
• Loi de Laplace–Gauss ou Loi Normale 𝓝(𝒎, 𝝈𝟐 ) ∶
 Fonction densité de probabilité ∶
 Espérance mathématique∶
 Variance∶
 Coefficient d'asymétrie∶
 Coefficient d'aplatissement (kurtosis) :

360 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
 Fonction génératrice des moments∶
 Propriétés ∶
 Approximation d’une loi Binomiale et de la loi de Poisson par une loi Normale ∶
 Utilisation∶
• Loi Normale centrée réduite ou Loi Normale standard N(0, 1) ∶
 Fonction densité de probabilité ∶
 Espérance mathématique∶
 Variance∶
 Médiane∶
 Fonction génératrice des moments∶
 Propriétés de la Fonction de répartition ∶
• Loi Log–normale LN(m, σ2 ) ∶
 Fonction densité de probabilité ∶
 Fonction de répartition∶
 Espérance mathématique∶
 Variance∶
 Utilisation∶
• Lois de Cauchy (ou Loi de Lorentz) C(0 , a) ∶
 Fonction densité de probabilité ∶
 Fonction de répartition∶
 Espérance mathématique∶
 Variance∶
 Fonction génératrice des moments∶
 Remarques∶
• Loi de Pareto P(c , α) ∶
 Fonction densité de probabilité ∶
 Fonction de répartition∶
 Espérance mathématique∶
 Variance∶
 Fonction génératrice des moments∶
 Propriétés d'une loi à queue longue (ou longue traîne ) ∶
• Loi de Laplace L( λ) ∶
 Fonction densité de probabilité ∶
 Fonction de répartition∶
 Espérance mathématique∶
 Variance∶
 Fonction génératrice des moments∶
• Loi de Gumbel G(μ , β) ∶
 Fonction densité de probabilité ∶
 Fonction de répartition∶
 Espérance mathématique∶
 Variance∶
• Loi Gamma Γ(α, β) , α > 0 , β > 0 ∶
 Fonction densité de probabilité ∶
 Propriétés de la fonction Gamma d'Euler Γ α ()∶
 Espérance mathématique∶
 Variance∶
 Coefficient d'asymétrie∶
 Coefficient d'aplatissement (kurtosis) :
 Fonction génératrice des moments∶
 Propriétés de la loi Gamma ∶
• Loi bêta B(α, β), α > 0 et β > 0 ∶
 Fonction densité de probabilité ∶
 Propriétés de la fonction Bêta B(α, β) ∶
 Espérance mathématique∶
 Variance∶
 Propriétés de la loi Bêta ∶
• Loi de Pearson ou khi‑deux à n degrés de liberté χ2 (k) , k ∈ N * ∶
 Fonction densité de probabilité ∶
 Espérance mathématique∶
 Variance∶
 Coefficient d'asymétrie∶
 Coefficient d'aplatissement (kurtosis) :
 Fonction génératrice des moments∶
 Propriétés de la loi khi‑deux à n degrés de liberté χ2(k) ∶
• Loi de Student à k degrés de liberté T(k), k ∈ N * ∶
 Fonction densité de probabilité ∶
 Espérance mathématique∶
 Variance∶
 Coefficient d'asymétrie∶
 Coefficient d'aplatissement (kurtosis) :
 Propriétés de la loi de Student à k degrés de liberté T(k), k ∈ N* ∶
• Loi de Fisher‑Snedecor à m et n degrés de liberté F(m, n) , m, n ∈ N * ∶
 Fonction densité de probabilité ∶
 Espérance mathématique∶

361 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
 Variance∶
 Propriétés de la loi de Fisher‑Snedecor à m et n degrés de liberté F(m, n) , m, n ∈ N* ∶
Modes de convergence…..………………………………………………………………………………….. 299
• Inégalité de Markov∶
• Inégalité de Bienaymé‑Tchebychev∶
• Convergence presque sûre∶
a. Définition 1∶
b. Définition 2∶
• Convergence en probabilité ∶
a. Théorème de Slutsky∶
• Convergence en moyenne quadratique∶
• Convergence en loi∶
h. Théorème de Moivre‑Laplace∶
• Loi faible des grands nombres∶
• Loi forte des grands nombres∶
• Théorème central limite∶
Vecteurs aléatoires continues..…..…………………………………………………………………………... 303
• Vecteurs aléatoires ∶
• Densité conjointe∶
• Densités marginales∶
 Remarques ∶
 Attention ∶
• Indépendance de deux v.a.continues ∶
• Indépendance d’une famille finie de v.a.∶
• Indépendance d’une suite de v.a.∶
• Indépendance de fonctions des variables aléatoires continue∶
• Somme de deux v.a.∶
• Fonction de répartition d'un couple de v.a.continues∶
• Fonctions de répartitions marginales∶
• Conditionnement ou distributions conditionnelles∶
• L’espérance mathématique d’une fonction φ de (X,Y) ∶
• Espérance conditionnelle∶
f. Inégalité de Cauchy Schwarz∶
• Variance‑Covariance ∶
 Propriétés ∶
n. Variances Conditionnelles ∶
o. Inégalité de Cauchy Schwarz :
p. Matrice de Covariance∶
 Cas des variables indépendantes ∶
 Remarques:
• Coefficient de corrélation linéaire ∶
 Propriétés ∶
 Remarques:
• Fonction génératrice des moments ∶
 Propriétés ∶
• Maximum et minimum de variables aléatoires continues et indépendantes ∶
Vecteurs Gaussiens...…..…………………………………………………………………………………… 309
• Exemple fondamental∶
• Définition∶
a. Remarques ∶
b. Proposition 1 :
c. Proposition 2 ∶
d. Transformation linéaire d’un vecteur gaussien ∶
e. Vecteur gaussien standard∶
• Indépendance de variables gaussiennes∶
a. Proposition 1 ∶
b. Corollaire :
c. Proposition 2 ∶
ème
Exercice 1 : (I.FI.D XXVII PROMO JUILLET 2007)................................................................................................312
Exercice 2 : (I.FI.D XXVIème PROMO JUILLET 2006) ……………………………………………………………………………………………...313
Exercice 3 : (I.FI.D XXXVIème PROMOTION (BANQUE) AOÛT 2016) ……………………………………………………………………..315

362 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba
BEN AHMED MOHSEN Téléphone: (+216) 97 619191 / 54 619191
[email protected] https://web.facebook.com/OMEGACENTER2014
Exercice 4 : (I.FI.D XXIVème PROMO JUILLET 2004) ……………………………………………………………………………………………...317
Exercice 5 : (I.FI.D XXIXème PROMO JUILLET 2009) ……………………………………………………………………………………………...322
Exercice 6 : (I.FI.D XXXVIIème PROMOTION (BANQUE) AOÛT 2017) …………………………………………………………………….325
Exercice 7 : (I.FI.D XXXVIIème PROMOTION (ASSURANCE) Avril 2018) …………………………………………………………………327
Exercice 8 : (I.FI.D XXXVIIIème PROMO ASSURANCE Septembre 2020)………………………………………………………………..329
Exercice 9 (Indépendance de variables aléatoires gaussiennes et corrélation) : ………………………………………………………………………………332
Exercice 10 (Transformations de variables-Inégalité de Jensen) : ……………………………………………………………………………………………..333
Exercice 11 (Corrélation entre variables la loi normale centrée réduite) : …………………………………………………………………………………….334
Exercice 12 (Transformations de variables-Loi Exponentielle) : ……………………………………………………………………………………………….335
Exercice 13 (Transformations de variables-Loi Normale centrée-réduite) : ……………………………………………………………………………………335
Exercice 14 (Transformations de variables- Loi khi‑deux à un degré de liberté) : …………………………………………………………………………….336
Exercice 15 (Transformations de variables-Loi Uniforme continue-Loi Exponentielle) : ………………………………………………………………………337
Exercice 16 (Transformations de variables-Loi de Weibull-Loi Exponentielle) : ……………………………………………………………………………….337
Exercice 17 (Transformations de variables-Loi de Cauchy) : ……………………………………………………………………………………………………338
Exercice 18 (Répartition des richesses-Loi de Pareto) : …………………………………………………………………………………………………………339
Exercice 19 (Maximum de variables aléatoires continues et indépendantes) : ……………………………………………………………………………….342
Exercice 20 (Inégalité de Tchebychev) : ………………………………………………………………………………………………………………………….343
Exercice 21 (Transformations de variables-Loi log‐normale-Modes de convergence) : ………………………………………………………………………344
Exercice 22 (Minimum de variables aléatoires continues et indépendantes-Modes de convergence) :……………………………………………………..347
Exercice 23 (Variables de Bernoulli-Modes de convergence) : …………………………………………………………………………………………………350
Exercice 24 (Taux de panne- Loi de Weibull- Loi des extrêmes- Loi exponentielle) : ………………………………………………………………………….353
Exercice 25 (Variable tronquée) : ………………………………………………………………………………………………………………………………..354
Exercice 26 (Théorème de Fisher-Cochran) : ..…………………………………………………………………………………………………………………..357

363 40ème Promotion Banque 2020/Axe⑤

Adresse : 2, Rue Ibn Arafa, av de l'environnement Manouba Centre, Manouba