[go: up one dir, main page]

Skip to content

Versions of hybrid pso algorithms for engineering optimization

Notifications You must be signed in to change notification settings

DeriZSY/hybrid_mopso

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

37 Commits
 
 
 
 
 
 

Repository files navigation

HybridPSODevelopment

Update and Notification :


All the work in this rep is finished as a research assistant of Advanced Manufacturing Group of Department of Mechanical and Energy Engineering, Southern University of Science and Technology,cooperating with PhD candidate Zhanying CHEN , and under the supervised of Prof. Xuekun LI. Both of my supervisor and cooperator are from Institude of Manufacturing Engineering, Tsinghua University.

Reference literature will be supplied with review later. (when I'm out of endless exams and projects)

ALL COMPYRIGHTS RESERVED


Depiction


This project aimed to develop extended PSO algorithm that is of praticle value for optimization problems in real engineering.

Features of the algorithm:

1. Multple objectives

Two sets of codes, uses relatively density measure and superposition method when select leader for the searm is provided.

2. Multiple constraints

In all codes provided, multiple constraints are handled with feasible area method (which is proved to be a more efficient method compared to basic projection method ).

3. Setable weights on different objectives

Although codes selecting leader with density measure is provided, all codes select a final global optima relys on Grid Index, which is essentially superposition of objective functions. The weights attached to different objectives are setable.

4. Produces a unique global optima together with Pareto Front

The codes in the pack are set to graph the Pareto Front, as well as using a blue square to suggest position of the global optima. Codes which can be used to graph the process of all particles together with Pareto Front is also reserved (and commented). However there may be particles whose values of objective functions might include imaginary numbers, so it's not suggested to be used.

Codes in the pack

MOPSO(density meausre method)

Matlab codes of MOPSO algorithm uses density measure.

mopsoDensityMeasure

Basic MOPSO using density measure.

mopsoVariantParameter

MOPSO using density measure, and improved with vairnt weight and learning factor (c1 and c2) duirng runtime. (I've examined this two algorithms in several different ways, but end up failling to decide which one performs better).

MOPSO(Superposition Method)

MOPSO using superposition method and basic, non-variant coeeficient.

Reference

Very original version of codes comes from : https://www.mathworks.com/matlabcentral/fileexchange/52870-multi-objective-particle-swarm-optimization--mopso-

More detaied description will be presented in technical document (which by now, may still reamianed unfinished).

You are welcome to contact me for any problems at: derizsy@foxmail.com

About

Versions of hybrid pso algorithms for engineering optimization

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages