Scribd
Upload
21 views2 pages

M N×T N T S S T S S T: Li and Zhou, New Paradigm: EEPSE Proposal

The document outlines the application of Particle Swarm Optimization (PSO) in Time-of-Use (TOU) based manufacturing systems, detailing a step-by-step process. It includes generating a population of particles, evaluating their fitness using a penalty function, and updating their velocity and position. The process is repeated until a maximum number of iterations is reached, with specific adjustments made for binary decision variables.

Uploaded by

kgkcjkkqq7
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
21 views2 pages

M N×T N T S S T S S T: Li and Zhou, New Paradigm: EEPSE Proposal

The document outlines the application of Particle Swarm Optimization (PSO) in Time-of-Use (TOU) based manufacturing systems, detailing a step-by-step process. It includes generating a population of particles, evaluating their fitness using a penalty function, and updating their velocity and position. The process is repeated until a maximum number of iterations is reached, with specific adjustments made for binary decision variables.

Uploaded by

kgkcjkkqq7
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

According to the knowledge of Chapter 3.

3, PSO is suitable for solving non-linear and


non-differentiable problems in high dimension space. In this section, the application of PSO in
TOU based manufacturing system will be explained step by step.

Step 1: Generate the population and initialization.

A population of particles need to be generated. The number of particles in the population,


which is known as the population size M , should be decided. In the model, the particles of
optimal problem are N ×T dimension, which the position matrix of the particlescan be expressed
by the control signal si (t ) of the manufacturing system throughout the planning horizon with N
machine and T intervals.

[ ]
s 1 (1) ⋯ s1 (T )
X= ⋮ ⋱ ⋮
s N (1) ⋯ s N (T ) (6-1)

Because the decision variable si (t ) is binary and not continuous. The initial value of all
elements of particles position is randomly selected from the set {0 ,1 }. The elements of the initial
velocity of each particle for particles are randomly generated from the set {−1 , 0 ,1 }.

Step 2: Evaluate the fitness function of each particle and update the gbest and pbest.

Since the standard PSO is only suitable for unconstrained optimization problems, the
penalty function will be used to establish the fitness function, which can be expressed as,

T N−1 T N−1
min f ( x ) +σ 1 [min ⁡(TP−TA ,0)] +σ 2 ∑ ∑ [min ⁡(C i−c i ,0)] +σ 3 ∑ ∑ [min ⁡(c i , 0)] 2
2 2

t =1 i=1 t =1 i=1 (6-2)

Where σ 1, σ 2, and σ 3 is the penalty parameters, which are extrem large real numbers.

gbest and pbest will be find according to the location of the particles and the fitness
function. The results will be stored for next iteration.

Step 3: Update the velocity and position

In basic PSO, the velocity and location of each particle are updated according to the
following equations,

v ¿ ( s+1)=ω v ¿ (s)+c 1 r 1 [ pbes t ¿ (s)−x¿ ]+c 2 r 2 [gbes t ¿ ( s)−x¿ ] (6-3)

Li and Zhou, New Paradigm: EEPSE Proposal Page 1


x ¿ (s+1)=x ¿ (s )+ v ¿ (s+ 1) (6-4)

Where v ¿ ( s) and x ¿ ( s) is the individual elements in the ith row and t th column of the velocity
and position matrix of the particle at the iteration s.

In our model because the decision variable si (t ) is binary, further steps are needed to
make the velocity and location be in the set {−1 , 0 ,1 } and {0 ,1 }.

v ¿ (s+1)=ω v ¿ (s)+c 1 r 1 [ pbes t ¿ (s)−x¿ ]+c 2 r 2 [gbes t ¿ (s)−x¿ ] (6-5)

{
1
1 , rand ( 0 ,1 ) <
x ¿ (s+1)= 1+e
−|x ¿ (s )+v ¿ (s +1)|

0 ,otherwise (6-6)

Step 4: Repeat steps 2 to 3 until the maximum number of iterations is reached.

Li and Zhou, New Paradigm: EEPSE Proposal Page 2

You might also like