PSO algorithm for Multi-Objective Optimization: Part 1- Development Concept

Adisorn Owatsiriwong, D.Eng.

The PSO framework for Multi-objective optimization (MOO) can be evolved from PSO code for single-objective problems. The key differences are in determining the personal best and global best solution. In our development, the personal best is the latest non-dominated position for each particle moving to the Pareto front. The global best is the best particle position that dominates all other solutions on the updated Pareto front.

The following diagram shows what the PSO-MOO algorithm looks like.

The function to determine if the set of objective functions [x] dominates the set of objective functions [y] is written as:

function [o] = dominate(x,y)

o = all(x<=y) && any(x<y) then

end function

Citation: OpenAI. (2024). ChatGPT [Large language model]. /g/g-5QhhdsfDj-diagrams-show-me