Global-Optimization-HAGCS

Introduction

A new global optimization algorithm developed by using gradient infromation, multiple search strategies, and parameters self-adaptation.

HAGCS is desgined for the objective functions whose gradient is readily available. Essentially, the HAGCS algorithm is based on the Cuckoo search [CS, 1] algorithm.

Installation

Download the required libraries: git clone https://github.com/stepbystep88/WZDOM

Set path: addpath(genpath("./WZDOM"));

Test

1. View the animation of optimization process (run testShowAnimation.m )

Compare CS with HAGCS on the Drop Wave test function:

Compare CS with HAGCS on the Drop Wave test function (bigger population size):

Compare CS with HAGCS on the Rastrigin test function:

Compare CS with HAGCS on the Rastrigin test function (smaller population size):

Compare CS with HAGCS on the Schwefel test function:

Compare CS with HAGCS on the Schwefel's P2.22 test function:

Compare CS with HAGCS on the Ackley test function:

2. To compare the performance of finding global minimum, convergence speed of different algorithms, please run testCompareMethodsFinal.m

References

[1] Xin-She Yang and Suash Deb. Cuckoo search via levy flights. In 2009 World Congress on Nature Biologically Inspired Computing (NaBIC), pages 210-214, Dec 2009.