Finite Differences
Finite Differences
Finite Differences
6.1 Introduction
For a function 𝑦 = 𝑓 𝑥 , finite differences refer to
changes in values of 𝑦 (dependent variable) for any
finite (equal or unequal) variation in 𝑥 (independent
variable).
In this chapter, we shall study various differencing
techniques for equal deviations in values of 𝑥 and
associated differencing operators; also their
applications will be extended for finding missing
values of a data and series summation.
6.2 Shift or Increment Operator (𝑬)
Shift (Increment) operator denoted by ‘𝐸’ operates on 𝑓(𝑥) as 𝐸𝑓 𝑥 = 𝑓 𝑥 +
Or 𝐸𝑦𝑥 = 𝑦𝑥+ , where ‘’ is the step height for equi-spaced data points.
Clearly effect of the shift operator 𝐸 is to shift the function value to the next
higher value 𝑓 𝑥 + or 𝑦𝑥+
Also 𝐸 2 𝑓 𝑥 = 𝐸 𝐸𝑓(𝑥) = 𝐸𝑓 𝑥 + = 𝑓 𝑥 + 2
∴ 𝐸 𝑛 𝑓 𝑥 = 𝑓 𝑥 + 𝑛
Moreover 𝐸 −1 𝑓 𝑥 = 𝑓 𝑥 − , where 𝐸 −1 is the inverse shift operator.
6.3 Differencing Operators
If 𝑦0 , 𝑦1 , 𝑦2 , … 𝑦𝑛 be the values of 𝑦 for corresponding values of 𝑥0 , 𝑥1 , 𝑥2 ,
… 𝑥𝑛 , then the differences of 𝑦 are defined by (𝑦1 − 𝑦0 ), (𝑦2 − 𝑦1 ), … , (𝑦𝑛 −
𝑦𝑛−1 ) , and are denoted by different operators discussed in this section.
6.3.1 Forward Difference Operator ∆
Forward difference operator ‘∆’ operates on 𝑦𝑥 as ∆𝑦𝑥 = 𝑦𝑥+1 − 𝑦𝑥
Or ∆𝑓 𝑥 = 𝑓 𝑥 + − 𝑓 𝑥 , where is the height of differencing.
∴ ∆𝑦0 = 𝑦1 − 𝑦0
∆𝑦1 = 𝑦2 − 𝑦1
⋮
∆𝑦𝑛 = 𝑦𝑛 +1 − 𝑦𝑛
Also ∆2 𝑦0 = ∆𝑦1 − ∆𝑦0 = 𝑦2 − 𝑦1 − 𝑦1 − 𝑦0 = 𝑦2 − 2𝑦1 + 𝑦0
⋮
∆ 𝑦0 = 𝑦𝑛 − 𝑛 𝐶1 𝑦𝑛−1 + 𝑛 𝐶2 𝑦𝑛−2 − ⋯ + −1 𝑛−1 𝑛 𝐶𝑛−1 𝑦1 + −1 𝑛 𝑦0
𝑛
Result 1: The 𝒏𝒕𝒉 differences of a polynomial of degree 'n' are constant and all
higher order differences are zero.
Proof: Consider the polynomial 𝑓(𝑥) of 𝑛𝑡 degree
𝑓(𝑥) = 𝑎0 𝑥 𝑛 + 𝑎1 𝑥 𝑛 −1 + 𝑎2 𝑥 𝑛 −2 + ⋯ + 𝑎𝑛 −1 𝑥 + 𝑎𝑛
First differences of the polynomial 𝑓(𝑥) are calculated as:
∆ 𝑓 𝑥 = 𝑓 𝑥 + − 𝑓(𝑥)
= 𝑎0 (𝑥 + )𝑛 − 𝑥 𝑛 + 𝑎1 (𝑥 + )𝑛−1 − 𝑥 𝑛 −1 + ⋯ + 𝑎𝑛−1 𝑥 + − 𝑥
= 𝑎0 𝑛 𝑥 𝑛 −1 + 𝑎1′ 𝑥 𝑛−1 + 𝑎2′ 𝑥 𝑛 −2 + ⋯ + 𝑎𝑛−1
′
+ 𝑎𝑛′
′ ′ ′ ′
where 𝑎1 , 𝑎2 ,… ,𝑎𝑛 −1 , 𝑎𝑛 are new constants
⇒ First difference of a polynomial of degree 𝑛 is a polynomial of degree (𝑛 − 1)
Similarly ∆2 𝑓 𝑥 = ∆ 𝑓 𝑥 + − ∆ 𝑓 𝑥
= 𝑎0 𝑛(𝑛 − 1)2 𝑥 𝑛−2 + 𝑎1′′ 𝑥 𝑛−3 + … + 𝑎𝑛′′
∴ Second difference of a polynomial of degree 𝑛 is a polynomial of degree (𝑛 − 2)
Repeating the above process ∆𝑛 𝑓 𝑥 = 𝑎0 𝑛(𝑛 − 1) … 2.1𝑛 𝑥 𝑛−𝑛
⇒ ∆𝑛 𝑓 𝑥 = 𝑎0 𝑛! 𝑛 which is a constant
∴ 𝑛𝑡 Difference of a polynomial of degree 𝑛 is a polynomial of degree zero.
Thus (𝑛 + 1)𝑡 and higher order differences of a polynomial of 𝑛𝑡 degree are all
zero.
The converse of above result is also true , i.e. if the 𝑛𝑡 difference of a
polynomial given at equally spaced points are constant then the function is
a polynomial of degree ‘𝑛’.
6.3.2 Backward Difference Operator 𝛁
Backward difference operator ‘ ∇ ’ operates on 𝑦𝑛 as ∇𝑦𝑛 = 𝑦𝑛 − 𝑦𝑛−1
∴ The differences (𝑦1 − 𝑦0 ) , (𝑦2 − 𝑦1 ) , … , (𝑦𝑛 − 𝑦𝑛−1 ) when denoted by
∇𝑦1 , ∇𝑦2 , … , ∇𝑦𝑛 are called first backward differences.
Also ∇2 𝑦𝑛 = ∇𝑦𝑛 − ∇𝑦𝑛 −1 , ∇3 𝑦𝑛 = ∇2 𝑦𝑛 − ∇2 𝑦𝑛−1 denote second and third
backward differences respectively.
Table 6.2 shows the backward differences of various orders.
Table 6.2 Backward Differences
𝒙 𝒚 𝛁 𝛁𝟐 𝛁𝟑 𝛁𝟒 𝛁𝟓
𝒙𝒐 𝑦𝑜
∇𝑦1
𝒙𝟏 𝑦1 ∇2 𝑦2
∇𝑦2 ∇3 𝑦3
𝒙𝟐 y2 ∇2 𝑦3 ∇4 𝑦4
∇𝑦3 ∇3 𝑦4 ∇5 𝑦5
2 4
𝒙𝟑 𝑦3 ∇ 𝑦4 ∇ 𝑦5
∇𝑦4 ∇3 𝑦5
𝒙𝟒 𝑦4 ∇2 𝑦5
∇𝑦5
𝒙𝟓 𝑦5
The arrow indicates the direction of differences from bottom to top. Differences in
each column notate difference of two adjoining consecutive entries of the previous
column, i.e. ∇𝑦1 = 𝑦1 − 𝑦𝑜 , ∇2 𝑦2 = ∇𝑦2 − ∇𝑦1 , … , ∇5 𝑦5 = ∇4 𝑦5 − ∇4 𝑦4 .
Relation between 𝛁 and 𝑬
∇ and 𝐸 are connected by the relation ∇ ≡ 1 − 𝐸 −1
Proof: we know that ∇𝑦𝑛 = 𝑦𝑛 − 𝑦𝑛 −1
= 𝑦𝑛 − 𝐸 −1 𝑦𝑛
⇒ ∇𝑦𝑛 = 1 − 𝐸 −1 𝑦𝑛
⇒ ∇ ≡ 1 − 𝐸 −1
6.3.3 Central Difference Operator 𝛅
Central difference operator ‘ δ ’ operates on 𝑦𝑛 as δ 𝑦𝑛 = 𝑦𝑛 +1 − 𝑦𝑛−1
2 2
∴ The differences (𝑦1 − 𝑦0 ) , (𝑦2 − 𝑦1 ) , … , (𝑦𝑛 − 𝑦𝑛−1 ) when denoted by
δ𝑦1 , δ𝑦3 , … , δ𝑦𝑛−1 are called first central differences.
2 2 2
⇒ 𝐸 𝑦 𝑥 = 𝑒 𝐷 𝑦(𝑥)
𝑑
∴ 𝐸 = 𝑒 𝐷 , 𝐷 ≡
𝑑𝑥
𝒅
Result 3: Relation between ∆ and 𝑫, where 𝑫 ≡
𝒅𝒙
We know that ∆ ≡ 𝐸 − 1
⇒ ∆ ≡ 𝑒 𝐷 − 1 ∵ 𝐸 = 𝑒 𝐷
𝒅
Result 4: Relation between 𝜵 and 𝑫, where 𝑫 ≡
𝒅𝒙
−1 −𝐷
We know that ∇ ≡ 1 − 𝐸 =1−𝑒 ∵ 𝐸 = 𝑒 𝐷
Result 5: Relation between ∆ and 𝜵
We know that 𝐸 ≡ 1 + ∆ ⋯①
Also 𝐸 −1 ≡ 1 − ∇
1
⇒ 𝐸≡ ⋯②
1−∇
1
⇒ 1 + ∆≡ From ① and ②
1−∇
1
⇒ ∆≡ −1
1−∇
∇
⇒ ∆≡
1−∇
Result 6: Relation between 𝛍 , 𝜹 and 𝑬
1 1 1
We have μ≡ 𝐸2 + 𝐸− 2
2
1 1
Also δ ≡ 𝐸 2 − 𝐸− 2
1 1 1 1 1
⇒ μδ ≡ 𝐸2 + 𝐸− 2 𝐸2 − 𝐸− 2
2
1
⇒ μδ ≡ 𝐸 − 𝐸− 1
2
Result 7: Relation between 𝛍 , 𝜹 , ∆ and ∇
1 1
We have μδ ≡ 𝐸 − 𝐸− 1 = 1 + ∆) − (1 − ∇
2 2
1
⇒ μδ ≡ ∆+∇
2
Result 8: ∆𝑛 𝑦𝑟 = ∇𝑛 𝑦𝑛 +𝑟
We have ∆𝑛 𝑦𝑟 = (𝐸 − 1)𝑛 𝑦𝑟 ∵ ∆= 𝐸 − 1
= 𝑦𝑛 +𝑟 − 𝑛 𝐶1 𝑦𝑛+𝑟−1 + 𝑛 𝐶2 𝑦𝑛+𝑟−2 − ⋯ + −1 𝑟 𝑦𝑟
= 𝐸 𝑛 − 𝑛 𝐶1 𝐸 𝑛−1 + 𝑛 𝐶2 𝐸 𝑛 −2 − ⋯ + −1 𝑛
𝑦𝑟
= 𝐸 𝑛 𝑦𝑟 − 𝑛 𝐶1 𝐸 𝑛−1 𝑦𝑟 + 𝑛 𝐶2 𝐸 𝑛−2 𝑦𝑟 − ⋯ + −1 𝑛 𝑦𝑟
= 𝑦𝑛 +𝑟 − 𝑛 𝐶1 𝑦𝑛+𝑟−1 + 𝑛 𝐶2 𝑦𝑛+𝑟−2 − ⋯ + −1 𝑛 𝑦𝑟
Also ∇𝑛 𝑦𝑛 +𝑟 = (1 − E −1 )𝑛 𝑦𝑛+𝑟 ∵ ∇ ≡ 1 − 𝐸 −1
= 1 − 𝑛 𝐶1 𝐸 −1 + 𝑛 𝐶2 𝐸 −2 − ⋯ + −1 𝑛 𝐸 −𝑛 𝑦𝑛+𝑟
= 𝑦𝑛 +𝑟 − 𝑛 𝐶1 𝑦𝑛+𝑟−1 + 𝑛 𝐶2 𝑦𝑛+𝑟−2 − ⋯ + −1 𝑛 𝑦𝑟
∴ ∆𝑛 𝑦𝑟 = ∇𝑛 𝑦𝑛 +𝑟
Example 1 Evaluate the following:
𝑥+1
i. ∆𝑒 𝑥 ii. ∆2 𝑒 𝑥 iii. ∆ 𝑡𝑎𝑛−1 𝑥 iv. ∆ v. ∆𝑓𝑘2 = 𝑓𝑘 + 𝑓𝑘+1 ∆𝑓𝑘
𝑥 2 −3𝑥+2
Solution: i. ∆𝑒 𝑥 = 𝑒 𝑥+ − 𝑒 𝑥 = 𝑒 𝑥 (𝑒 − 1)
∆𝑒 𝑥 = 𝑒 𝑥 (𝑒 − 1) , if = 1
ii. ∆2 𝑒 𝑥 = ∆(∆𝑒 𝑥 )
= ∆ 𝑒 𝑥 𝑒 − 1
= 𝑒 − 1 ∆𝑒 𝑥
= 𝑒 − 1 𝑒 𝑥 + − 𝑒 𝑥
= 𝑒 − 1 𝑒 𝑥 (𝑒 − 1)
= 𝑒 𝑥 (𝑒 − 1)2
iii. ∆𝑡𝑎𝑛−1 𝑥 = 𝑡𝑎𝑛−1 𝑥 + − 𝑡𝑎𝑛−1 𝑥
𝑥+−𝑥
= 𝑡𝑎𝑛−1
1+(𝑥+)𝑥
−1
= 𝑡𝑎𝑛
1+(𝑥+)𝑥
𝑥+1 𝑥+1
iv. ∆ = ∆
𝑥 2 −3𝑥+2 𝑥 −1 (𝑥−2)
−2 3 −2 3
=∆ + =∆ + ∆
𝑥−1 𝑥−2 𝑥−1 𝑥−2
1 1 1 1
= −2 − +3 −
𝑥+1−1 𝑥−1 𝑥+1−2 𝑥−2
1 1 1 1
= −2 − +3 −
𝑥 𝑥−1 𝑥−1 𝑥−2
(𝑥+4)
=−
𝑥 𝑥−1 (𝑥−2)
v. ∆𝑓𝑘2 = 𝑓𝑘+1
2
− 𝑓𝑘2 = 𝑓𝑘+1 + 𝑓𝑘 𝑓𝑘+1 − 𝑓𝑘 = 𝑓𝑘 + 𝑓𝑘+1 ∆𝑓𝑘
Example 2 Evaluate the following:
𝑥2
i. ∆𝑒 𝑥 log 2𝑥 ii. ∆
cos 2𝑥
= 𝑒 𝑥 𝑒 log 1 + + log 2𝑥 𝑒 − 1
𝑥
ii. Let 𝑓 𝑥 = 𝑥 2 and 𝑔 𝑥 = cos 2𝑥
𝑓 𝑥 𝑔 𝑥 ∆𝑓 𝑥 −𝑓(𝑥)∆𝑔 𝑥
We have ∆ =
𝑔 𝑥 𝑔 𝑥+ 𝑔 𝑥
𝐸−1 1−𝐸 –1
= –1
−
1−𝐸 𝐸−1
2
𝐸−1 2 − 1− 𝐸 –1
=
1−𝐸 –1 𝐸−1
𝐸 2 +1− 2𝐸 − 1+ 𝐸 –2 −2𝐸 –1
=
𝐸 + 𝐸 –1 −2
𝐸 2 −𝐸 −2 −2𝐸+2𝐸 –1
=
𝐸 + 𝐸 –1 −2
𝐸+𝐸 –1 𝐸−𝐸 –1 − 2 𝐸− 𝐸 –1
=
𝐸 + 𝐸 –1 −2
𝐸 − 𝐸 –1 𝐸 + 𝐸 –1 −2
=
𝐸 + 𝐸 –1 −2
= 𝐸 − 𝐸 −1 = R.H.S.
1 1
Example 6 Prove that 𝐸 = 1 + δ2 + δ 1 + δ2
2 4
1 1
Solution: R.H.S. = 1 + δ2 + δ 1 + δ2
2 4
1 1 1 2 1 1 1 1 1 2
=1+ 𝐸2 − 𝐸− 2 + 𝐸2 − 𝐸− 2 1+ 𝐸2 − 𝐸− 2
2 4
1 1
∵ δ ≡ 𝐸2 − 𝐸− 2
1 1 1 1
=1+ 𝐸 + 𝐸 −1 − 2 + 𝐸 2 − 𝐸 − 2 1+ 𝐸 + 𝐸 −1 − 2
2 4
1 1 1 1
=1+ 𝐸 + 𝐸 −1 − 2 + 𝐸 2 − 𝐸 − 2 𝐸 + 𝐸 −1 + 2
2 4
1 1 1 1 1 1 2
=1+ 𝐸 + 𝐸 −1 − 2 + 𝐸 2 − 𝐸 − 2 𝐸2 + 𝐸− 2
2 4
1 1 1 1 1 1
=1+ 𝐸 + 𝐸 −1 − 2 + 𝐸2 − 𝐸− 2 𝐸2 + 𝐸− 2
2 2
1 1
= 𝐸 + 𝐸 −1 + 𝐸 − 𝐸 −1 = 𝐸 = L.H.S.
2 2
1 1
Example 7 Prove that ∇ = − δ2 + δ 1 + δ2
2 4
1 1
Solution: R.H.S. = − δ2 + δ 1 + δ2
2 4
1 1 1 2 1 1 1 1 1 2
=− 𝐸2 − 𝐸− 2 + 𝐸2 − 𝐸− 2 1+ 𝐸2 − 𝐸− 2
2 4
1 1
∵ δ ≡ 𝐸2 − 𝐸− 2
1 1 1 1
=− 𝐸 + 𝐸 −1 − 2 + 𝐸 2 − 𝐸 − 2 1+ 𝐸 + 𝐸 −1 − 2
2 4
1 1 1 1
=− 𝐸 + 𝐸 −1 − 2 + 𝐸 2 − 𝐸 − 2 𝐸 + 𝐸 −1 + 2
2 4
1 1 1 1 1 1 2
=− 𝐸 + 𝐸 −1 − 2 + 𝐸 2 − 𝐸 − 2 𝐸2 + 𝐸− 2
2 4
1 1 1 1 1 1
=− 𝐸 + 𝐸 −1 − 2 + 𝐸2 − 𝐸− 2 𝐸2 + 𝐸− 2
2 2
1 1
=− 𝐸 + 𝐸 −1 − 2 + 𝐸 − 𝐸 −1 = 1 − 𝐸 −1 = ∇= L.H.S.
2 2
1
1 ∆ −
Example 8 Prove that (i) ∆ − ∇ = δ (ii) μ = 2
1+ δ2 = 1+ 1+∆ 2
4 2
1 1 2 1 1
Solution: (i) δ2 = 𝐸 2 − 𝐸 − 2 = 𝐸 + 𝐸 −1 − 2 ∵ δ ≡ 𝐸2 − 𝐸− 2
= 𝐸 − 1 − 1 − 𝐸 −1 = ∆ − ∇
∵ 𝐸 − 1 ≡ ∆ and 1 − 𝐸 −1 = ∆
1 1 1 1 2 1 1
(ii) 1 + δ2 = 1+ 𝐸2 − 𝐸− 2 ∵ δ ≡ 𝐸2 − 𝐸− 2
4 4
1
= 1+ 𝐸 + 𝐸 −1 − 2
4
1
= 𝐸 + 𝐸 −1 + 2
4
1 1 1 2
= 𝐸2 + 𝐸− 2
4
1 1 1 1 1 1
= 𝐸2 + 𝐸− 2 = μ ∵μ≡ 𝐸2 + 𝐸− 2
2 2
∆ 1 𝐸−1 1
− −
Also 1 + 1+∆ 2 = 1+ 1+𝐸−1 2
2 2
∵∆ ≡𝐸−1
1
𝐸+1
= 𝐸− 2
2
1 1 1
= 𝐸− 2 + 𝐸2 = μ
2
1
𝛿 2𝑟
Example 9 Prove that (i) ∆ ≡ 𝐸∇ ≡ ∇E = δ𝐸 2 (ii) Er = μ +
2
Solution: (i) 𝐸∇ = E 1 − 𝐸 − 1 = 𝐸 − 1 = ∆ ∵ ∇ ≡ 1 − 𝐸 −1
∇E = 1 − 𝐸 − 1 𝐸 = 𝐸 − 1 = ∆
1 1 1 1 1 1
δ𝐸 2 = 𝐸 2 − 𝐸 − 2 𝐸 2 = 𝐸 − 1 = ∆ ∵ δ ≡ 𝐸2 − 𝐸− 2
2𝑟
𝛿 2𝑟 1 1
−
1 1 1
−
1
(ii) R.H.S. = μ + = 𝐸 +𝐸2 2 + 𝐸 −𝐸
2 2
2 2 2
1 1 1 1 1
∵μ≡ 𝐸 2 + 𝐸 − 2 and δ ≡ 𝐸 2 − 𝐸 − 2
2
2𝑟
1 1 1 2𝑟
= 2𝐸 2 = 𝐸2 = Er = L.H.S
2
1
Example 10 Prove that (i) 𝐷 ≡ 𝑙𝑜𝑔 𝐸 (iii) 𝐷 ≡ 𝑙𝑜𝑔 1 + ∆ ≡ −log
(1 − ∇)
7
(iii) ∇2 ≡ 2 𝐷 2 − 3 𝐷 3 + 4 𝐷 4 + ⋯
12
Remark: In order to prove any relation, we can express the operators (∆ ,∇,𝛿) in
terms of fundamental operator 𝐸.
Example 11 Form the forward difference table for the function
𝑓 𝑥 = 𝑥 3 − 2𝑥 2 − 3𝑥 − 1 for 𝑥 = 0, 1, 2, 3, 4.
Hence or otherwise find ∆3 𝑓 𝑥 , also show that ∆4 𝑓 𝑥 = 0
Solution: 𝑓 0 = −1, 𝑓 1 = −5, 𝑓 2 = −7, 𝑓 3 = −1, 𝑓 4 = 19
Constructing the forward difference table:
𝒙 𝒇(𝒙) ∆ ∆𝟐 ∆𝟑 ∆𝟒
𝟎 −1
−4
1 −5 2
−2 6
𝟐 −7 8 0
6 6
𝟑 −1 14
20
𝟒 19
From the table, we see that ∆3 𝑓 𝑥 = 6 and ∆4 𝑓 𝑥 = 0
Note: Using the formula ∆𝑛 𝑓 𝑥 = 𝑎0 𝑛! 𝑛 , ∆3 𝑓 𝑥 = 1.3!. 1𝑛 = 6
Also ∆𝑛+1 𝑓 𝑥 = 0 for a polynomial of degree 𝑛, ∴ ∆4 𝑓 𝑥 = 0
Example 12 If for a polynomial, five observations are recorded as: 𝑦0 = −8,
𝑦1 = −6, 𝑦2 = 22, 𝑦3 = 148, 𝑦4 = 492, find 𝑦5 .
Solution: 𝑦5 = 𝐸 5 𝑦0 = (1 + ∆)5 𝑦0 ∵𝐸 ≡1+∆
= 𝑦0 + 𝐶1 ∆𝑦0 + 𝐶2 ∆ 𝑦0 + 5 𝐶3 ∆3 𝑦0 +
5 5 2 5
𝐶4 ∆4 𝑦0 + ∆5 𝑦0 … ①
Constructing the forward difference table:
𝒙 𝒚 ∆ ∆𝟐 ∆𝟑 ∆𝟒
𝒙𝟎 −8
2
𝒙𝟏 −6 26
28 72
𝒙𝟐 22 98 48
126 120
𝒙𝟑 148 218
344
𝒙𝟒 492
From table ∆𝑦0 = 2, ∆2 𝑦0 = 26 , ∆3 𝑦0 = 72 , ∆3 𝑦0 = 48 …②
⇒ 𝑦5 = −8 + 5 2 + 10 26 + 10 72 + 5 48 = 1222 using ② in ①
6.4 Missing values of Data
Missing data or missing values occur when an observation is missing for a
particular variable in a data sample. Concept of finite differences can help to locate
the requisite value using known concepts of curve fitting.
To determine the equation of a line (equation of degree one), we need at least two
given points. Similarly to trace a parabola (equation of degree two), at least three
points are imperative. Thus we essentially require 𝑛 + 1 known observations to
determine a polynomial of 𝑛𝑡 degree.
To find missing values of data using finite differences, we presume the degree of
the polynomial by the number of known observations and use the result
∆𝑛+1 𝑓 𝑥 = 0 for a polynomial of degree 𝑛.
Example 13 Use the concept of missing data to find 𝑦5 if 𝑦0 = −8, 𝑦1 = −6,
𝑦2 = 22, 𝑦3 = 148, 𝑦4 = 492
Solution: Constructing the forward difference table taking 𝑦5 as missing value
𝒙 𝒚 ∆ ∆𝟐 ∆𝟑 ∆𝟒 ∆𝟓
𝒙𝒐 −8
2
𝒙𝟏 −6 26
28 72
𝒙𝟐 22 98 48
126 120 𝑦5 − 1222
𝒙𝟑 148 218 𝑦5 − 1174
344 𝑦5 − 1054
𝒙𝟒 492 𝑦5 − 836
𝑦5 − 492
𝒙𝟓 𝑦5
Since 5 observations are known, let us assume that the polynomial represented by
given data is of 4𝑡 degree. ∴ ∆5 𝑦 = 0 ⇒ 𝑦5 − 1222 = 0 or 𝑦5 = 1222
Example 14 Find the missing values in the following table
𝒙 𝟎 𝟓 𝟏𝟎 𝟏𝟓 𝟐𝟎 𝟐𝟓
𝒇(𝒙) 6 ? 13 17 22 ?
Solution: Since there are 4 known values of 𝑓 𝑥 in the given data, let us assume
the polynomial represented by the given data to be of 3𝑟𝑑 degree.
Constructing the forward difference table taking missing values as 𝑎 and 𝑏.
𝒙 𝒚 ∆ ∆𝟐 ∆𝟑 ∆𝟒
𝟎 6
𝑎−6
𝟓 𝑎 19 − 2𝑎
13 − 𝑎 3𝑎 − 28
𝟏𝟎 13 𝑎−9 38 − 4𝑎
4 10 − 𝑎
15 17 1 𝑎 + 𝑏 − 38
5 𝑏 − 28
2𝟎 22 𝑏 − 27
𝑏 − 22
25 𝑏
Since the polynomial represented by the given data is considered to be of
3𝑟𝑑 degree, 4𝑡ℎ and higher order differences are zero i.e. ∆4 𝑦 = 0
∴ 38 − 4𝑎 = 0 and 𝑎 + 𝑏 − 38 = 0
Solving these two equations, we get 𝑎 = 9.5 𝑏 = 28.5
6.5 Finding Differences Using Factorial Notation
We can conveniently find the forward differences of a polynomial using factorial
notation.
6.5.1 Factorial Notation of a Polynomial
A product of the form 𝑥 𝑥 − 1 𝑥 − 2 … 𝑥 − 𝑟 + 1 is called a factorial
polynomial and is denoted by 𝑥 𝑟
∴ 𝑥 =𝑥
𝑥 2 = 𝑥(𝑥 − 1)
𝑥 3 = 𝑥 𝑥 − 1 (𝑥 − 2)
⋮
𝑥 𝑛 =𝑥 𝑥−1 𝑥−2 … 𝑥−𝑛+1
In case, the interval of differencing is , then
𝑛
𝑥 = 𝑥 𝑥 − 𝑥 − … 𝑥 − 𝑛– 1
The results of differencing 𝑥 𝑟 are analogous to that differentiating 𝑥 𝑟
∴ ∆ 𝑥 𝑛 = 𝑛 𝑥 𝑛−1
∆2 𝑥 𝑛 = 𝑛(𝑛 − 1) 𝑥 𝑛−2
∆3 𝑥 𝑛 = 𝑛 𝑛 − 1 (𝑛 − 2) 𝑥 𝑛−3
⋮
𝑛 𝑛
∆ 𝑥 = 𝑛 𝑛 − 1 𝑛 − 2 … 3.2.1 = 𝑛!
∆𝑛+1 𝑥 𝑛 = 0
1 𝑥 2 1 2 𝑥 3
Also 𝑥 = , 𝑥 = and so on
∆ 2 ∆ 3
1 1 𝑥 2 𝑥 3
𝑥 = =
∆2 ∆ 2 6
⋮
Remark:
i. Every polynomial of degree 𝑛 can be expressed as a factorial
polynomial of the same degree and vice-versa.
ii. The coefficient of highest power of 𝑥 and also the constant term
remains unchanged while transforming a polynomial to factorial
notation.
Example15 Express the polynomial 2𝑥 2 + 3𝑥 + 1 in factorial notation.
Solution: 2𝑥 2 − 3𝑥 + 1 = 2𝑥 2 − 2𝑥 + 5𝑥 + 1
= 2𝑥 𝑥 − 1 + 5𝑥 + 1
=2 𝑥 2+5 𝑥 +1
Example16 Express the polynomial 2𝑥 3 − 𝑥 2 + 3𝑥 − 4 in factorial notation.
Solution: 2𝑥 3 − 𝑥 2 + 3𝑥 − 4 = 2 𝑥 3 + 𝐴 𝑥 2 + 𝐵 𝑥 − 4
Using remarks i and ii
= 2𝑥 𝑥 − 1 𝑥 − 2 + 𝐴𝑥 𝑥 − 1 + 𝐵𝑥 − 4
= 2𝑥 3 + 𝐴 − 6 𝑥 2 + −𝐴 + 𝐵 + 4 𝑥 − 4
Comparing the coefficients on both sides
A − 6 = −1, −𝐴 + B + 4 = 3
⇒ A = 5, B = 4
∴ 2𝑥 3 − 𝑥 2 + 3𝑥 − 4 = 2 𝑥 3 + 5 𝑥 2 + 4 𝑥 − 4
We can also find factorial polynomial using synthetic division as shown:
Coefficients 𝐴 and 𝐵 can be found as remainders under 𝑥 2 and 𝑥 columns
𝑥3 𝑥2 𝑥
1 2 –1 3 –4
– 2 1
2 2 1 4=𝐵
– 4
2 5=A
Example 17 Find ∆3 𝑓 𝑥 for the polynomial 𝑓 𝑥 = 𝑥 3 − 2𝑥 2 − 3𝑥 − 1
Also show that ∆4 𝑓 𝑥 = 0
Solution: Finding factorial polynomial of 𝑓 𝑥 as shown:
Let 𝑥 3 − 2𝑥 2 − 3𝑥 − 1 = 𝑥 3 + 𝐴 𝑥 2 + 𝐵 𝑥 − 1
Coefficients 𝐴 and 𝐵 can be found as remainders under 𝑥 2 and 𝑥 columns
𝑥3 𝑥2 𝑥
1 1 –2 –3 –1
– 1 –1
2 1 –1 –4 = 𝐵
– 2
1 1=A
∴ 𝑓 𝑥 = 𝑥 3 − 2𝑥 2 − 3𝑥 − 1 = 𝑥 3 + 𝑥 2 − 4 𝑥 − 1
∆3 𝑓 𝑥 = ∆3 𝑥 3 + 𝑥 2 − 4 𝑥 − 1
= 3! + 0 = 6 ∵ ∆𝑛 𝑥 𝑛 = 𝑛! and ∆𝑛+1 𝑥 𝑛 = 0
Also ∆4 𝑓 𝑥 = ∆4 𝑥 3 + 𝑥 2 − 4 𝑥 − 1 = 0
Note: Results obtained are same as in Example 11, where we have used
forward difference table to compute the differences.
Example 18: Obtain the function whose first difference is 8𝑥 3 − 3𝑥 2 + 3𝑥 − 1
Solution: Let 𝑓 𝑥 be the function whose first difference is 8𝑥 3 − 3𝑥 2 + 3𝑥 − 1
⇒ ∆𝑓 𝑥 = 8𝑥3 − 3𝑥2 + 3𝑥 − 1
Let 8𝑥 3 − 3𝑥 2 + 3𝑥 − 1 = 8 𝑥 3 + 𝐴 𝑥 2 + 𝐵 𝑥 − 1
Coefficients 𝐴 and 𝐵 can be found as remainders under 𝑥 2 and 𝑥 columns
𝑥3 𝑥2 𝑥
1 8 –3 3 –1
– 8 5
2 8 5 8=𝐵
– 16
8 21 = A
∴ ∆𝑓 𝑥 = 8𝑥3 − 3𝑥2 + 3𝑥 − 1 = 8 𝑥 3 + 21 𝑥 2 + 8 𝑥 − 1
1 3 2
𝑓 𝑥 = 8𝑥 + 21 𝑥 +8 𝑥 −1
∆
8𝑥 4 21 𝑥 3 8𝑥 2 1 𝑥 2 1 2 𝑥 3
= + + − 𝑥 ∵ 𝑥 = , 𝑥 = ,…
4 3 2 ∆ 2 ∆ 3
= 2 𝑥 4 + 7 𝑥 3 + 4 𝑥 2 − [𝑥]
= 2𝑥 𝑥 − 1 𝑥 − 2 𝑥 − 3 + 7𝑥 𝑥 − 1 𝑥 − 2 + 4𝑥 𝑥 − 1 − 𝑥
=𝑥 2 𝑥−1 𝑥−2 𝑥−3 +7 𝑥−1 𝑥−2 +4 𝑥−1 −1
= 𝑥 2𝑥 3 − 5𝑥 2 + 5𝑥 − 3 = 2𝑥 4 − 5𝑥 3 + 5𝑥 2 − 3𝑥
⇒ 𝑓 𝑥 = 2𝑥4 − 5𝑥3 + 5𝑥2 − 3𝑥
6.6 Series Summation Using Finite Differences
The method of finite differences may be used to find sum of a given series by
applying the following algorithm:
1. Let the series be represented by 𝑢0 , 𝑢1 , 𝑢2 , 𝑢3 , …
2. Use the relation 𝑢𝑟 = 𝐸 𝑟 𝑢0 to introduce the operator 𝐸 in the series.
3. Replace 𝐸 by ∆ by substituting 𝐸 ≡ 1 + ∆ and find the sum the series by
any of the applicable methods like sum of a G.P., exponential or logarithmic
series or by binomial expansion and operate term by term on 𝑢0 to find the
required sum.
Example 19 Prove the following using finite differences:
𝑥 𝑥2 ∆ 𝑢0 ∆2 𝑢 0
i. 𝑢0 + 𝑢1 + 𝑢2 + ⋯ = 𝑒 𝑥 𝑢0 + 𝑥 + 𝑥2 +⋯
1! 2! 1! 2!
1 1 1
ii. 𝑢0 − 𝑢1 + 𝑢2 − 𝑢3 + ⋯ = 𝑢0 − ∆ 𝑢0 + ∆2 𝑢0 − ⋯
2 4 8
𝑥 𝑥2 𝑥 𝑥2
Solution: i. 𝑢0 + 𝑢1 + 𝑢2 + ⋯ = 𝑢0 + 𝐸𝑢0 + 𝐸 2 𝑢0 + ⋯
1! 2! 1! 2!
𝑥𝐸 𝑥 2 𝐸2
= 1+ + 0 𝑢 + ⋯ 𝑢0
1! 2!
= 𝑒 𝑥𝐸 𝑢0 = 𝑒 𝑥(1+∆)
𝑢0
= 𝑒 𝑥 𝑒 𝑥∆ 𝑢0
𝑥∆ 𝑥 2 ∆2
= 𝑒𝑥 1 + + + ⋯ 𝑢0
1! 2!
∆ 𝑢0 ∆2 𝑢 0
= 𝑒 𝑥 𝑢0 + 𝑥 + 𝑥2 +⋯
1! 2!
ii. 𝑢0 − 𝑢1 + 𝑢2 − 𝑢3 + ⋯ = 𝑢0 − 𝐸𝑢0 + 𝐸2 𝑢0 − 𝐸 3 𝑢0 + ⋯
= 1 − 𝐸 + 𝐸 2 − 𝐸 3 + ⋯ 𝑢0
−1
= 1+𝐸 𝑢0
−1
= 2+∆ 𝑢0
∆ −1
= 2−1 1 + 𝑢0
2
1 ∆ ∆2
= 1− + − ⋯ 𝑢0
2 2 4
1 1 1
= 𝑢0 − ∆ 𝑢0 + ∆2 𝑢0 − ⋯
2 4 8
(1+∆)𝑛 −1
⇒𝑆= 𝑢0
(1+∆)−1
1 𝑛 (𝑛 −1) 𝑛 𝑛 −1 (𝑛 −2)
= 1 + 𝑛∆ + ∆2 + ∆3 + ⋯ − 1 𝑢0
∆ 2! 3!
𝑛 (𝑛 −1) 𝑛 𝑛 −1 (𝑛−2)
= 𝑛𝑢0 + ∆𝑢0 + ∆2 𝑢0 + ⋯
2! 3!
Now 𝑢0 = 12 = 1
∆𝑢0 = 𝑢1 − 𝑢0 = 22 − 12 = 3
∆2 𝑢0 = ∆𝑢1 − ∆𝑢0 = 𝑢2 − 2𝑢1 + 𝑢0 = 32 − 2( 22 ) + 12 = 2
∆3 𝑢0 , ∆4 𝑢0 … are all zero as given series is an expression of degree 2
𝑛 (𝑛 −1) 𝑛 𝑛 −1 (𝑛 −2)
∴ 𝑆=𝑛+ 3 + 2 +0
2! 3!
3𝑛 𝑛 −1 𝑛 𝑛 −1 𝑛 −2
=𝑛+ +
2 3
1
= 6𝑛 + 9𝑛 𝑛 − 1 + 2𝑛 𝑛 − 1 𝑛 − 2
6
1
= 𝑛 6 + 9𝑛 − 9 + 2𝑛2 − 6𝑛 + 4
6
1 1
= 𝑛 2𝑛2 + 3𝑛 + 1 = 𝑛 𝑛 + 1 (2𝑛 + 1)
6 6
𝑢0 𝑥∆𝑢 0 𝑥 2 ∆2𝑢 0
Example 21 Prove that 𝑢0 + 𝑢1 𝑥 + 𝑢2 𝑥 2 + ⋯ = + 2
+ +⋯
1−𝑥 (1−𝑥) (1−𝑥)3
= 1 + 𝑥𝐸 + 𝑥2 𝐸 2 + ⋯ 𝑢0
1 𝑎
= 𝑢0 ∵ 𝑆∞ =
1−𝑥𝐸 1−𝑟
1 1
= 𝑢0 = 𝑢0
1−𝑥(1+∆) (1−𝑥)−𝑥∆
1 1
=
1−𝑥 1− 𝑥∆
𝑢0
1−𝑥
1 𝑥∆ −1
=
1−𝑥
1 − 1−𝑥 𝑢0
1 𝑥∆ 𝑥2 ∆2
=
1−𝑥
1 + 1−𝑥 + (1−𝑥)2 + ⋯ 𝑢0
𝑢0 𝑥∆𝑢 0 𝑥 2 ∆2𝑢 0
= + + + ⋯ = R.H.S.
1−𝑥 (1−𝑥)2 (1−𝑥)3
Now to evaluate the series 1.2 + 2.3𝑥 + 3.4𝑥 2 + 4.5𝑥 3 + ⋯
Let 𝑢0 = 1.2 = 2, 𝑢1 = 2.3 = 6, 𝑢2 = 3.4 = 12, 𝑢3 = 4.5 = 20, …
Forming forward difference table to calculate the differences
𝒖 ∆ ∆𝟐 ∆𝟑 ∆𝟒
𝒖𝟎 = 𝟏. 𝟐 = 𝟐
4
𝒖𝟏 = 𝟐. 𝟑 = 𝟔 2
6 0
𝒖𝟐 = 𝟑. 𝟒 = 𝟏𝟐 2 0
8 0
𝒖𝟑 = 𝟒. 𝟓 = 𝟐𝟎 2
10
𝒖𝟒 = 𝟓. 𝟔 = 𝟑𝟎
𝑢0 𝑥∆𝑢 0 𝑥 2 ∆2𝑢 0
∴ 1.2 + 2.3𝑥 + 3.4𝑥 2 + 4.5𝑥 3 + ⋯ = + 2
+ +⋯
1−𝑥 (1−𝑥) (1−𝑥)3
2 4𝑥 2𝑥 2
= + + +0
1−𝑥 (1−𝑥)2 (1−𝑥)3
2
=
(1−𝑥)3
Exercise 6A
1. Express 𝑦4 in terms of successive forward differences.
2. Prove that ∆𝑛 𝑒 3𝑥+5 = 𝑒 3 − 1 𝑛 𝑒 3𝑥+5
5𝑥+12
3. Evaluate ∆2
𝑥 2 +5𝑥+6
4. If 𝑢0 = 3, 𝑢1 = 12, 𝑢2 = 81, 𝑢3 = 2000, 𝑢4 = 100, calculate ∆4 𝑢0 .
2+∆ 1
5. Prove that μ = = 1 + δ2
2 1+∆ 4
6. Find the missing value in the following table
𝑥 0 5 10 15 20 25
𝑦 6 10 - 17 - 31
7. Sum the series 13 , 23 , 33 ,…, 𝑛3 using finite differences.
Answers