Date of Award

11-27-2015

Publication Type

Master Thesis

Degree Name

M.Sc.

Department

Computer Science

Keywords

cultural algorithm, evolutionary algorithm, heritage, multi-population

Supervisor

Kobti, Ziad

Rights

info:eu-repo/semantics/openAccess

Abstract

Multi-Population Cultural Algorithms (MPCA) define a set of individuals that can be categorized as belonging to one of a set of populations. Not only reserved for Cultural Algorithms, the concept of Multi-Populations has been used in evolutionary algorithms to explore different search spaces or search for different goals simultaneously, with the capability of sharing knowledge with each other. The populations themselves can define specific goals or knowledge to use in the context of the problem. One limitation of MPCA is that an individual can only belong to one population at a time, which can restrict the potential and realism of the algorithm. This thesis proposes a novel approach to represent population usage called “Heritage,” which allows individuals to belong to multiple populations with weighted influence. Heritage-Dynamic Cultural Algorithm (HDCA) is used to test against different domains to examine the advantages and disadvantages of this approach.

Share

COinS