Seed disperser ant algorithm for optimization / Chang Wen Liang

The Seed Disperser Ant Algorithm (SDAA) is developed based on the evolution or expansion process of Seed Disperser Ant (Aphaenogaster senilis) colony. The genotype of every ant is represented in binary form as the variables. These binary variables are used to locally search for optimum solution. SDA...

Full description

Saved in:
Bibliographic Details
Main Author: Chang , Wen Liang
Format: Thesis
Published: 2018
Subjects:
Online Access:http://studentsrepo.um.edu.my/9377/1/Chang_Wen_Liang.jpg
http://studentsrepo.um.edu.my/9377/11/wen_liang.pdf
http://studentsrepo.um.edu.my/9377/
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The Seed Disperser Ant Algorithm (SDAA) is developed based on the evolution or expansion process of Seed Disperser Ant (Aphaenogaster senilis) colony. The genotype of every ant is represented in binary form as the variables. These binary variables are used to locally search for optimum solution. SDAA is developed using the concept of male ants performing nuptial flights to generate new superior colonies. The new colonies produce better male ants that repeat the nuptial flight cycle in following generation. New young queens are produced by the colony that migrates to establish new colonies after local optimum solution reached to start new local search. Nuptial flight and new young queens’ production aid in enhanced search exploitation and exploration respectively. This diversifies the search for global optimum. The classical benchmark problems and composite benchmark functions from Congress on Evolutionary Computation (CEC) 2005 special session is used for validate SDAA. Engineering optimization has become important in design problems to reduce error and faulty production as many constrained condition should be taken in to account before manufacturing. Also, data clustering has become popular in data mining in recent time due to data explosion. In this research, we applied SDAA to solve the constrained engineering problems and introduce an efficient data clustering algorithm which is hybrid of K-means and SDAA. The optimal results obtained for constrained engineering problems as well as data clustering are very promising in terms of quality of solutions and convergence speed of the algorithm.