The Development And Application Of Evolutionary Computation-Based Layered Encoding Cascade Optimization Model
Loading...
Date
2010-01
Authors
Neoh, Siew Chin
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Sains Malaysia
Abstract
In this thesis, the research on a generic evolutionary-based layered encoding cascade
optimization (LECO) model that is able to solve different kinds of optimization problems on
multi-decision, multi-resolution, interactive, hybridized and multi-objective is presented. In
the proposed model, particular attention is given to genetic algorithm (GA) and particle
swarm optimization (PSO) in the development of evolutionary-based search mechanism.
The foundation of the proposed model is the development of layered encoding representation
structure that integrates the ideas of divide and conquer strategy, schema theorem, and
hierarchical concepts of semantic network and frame-based representation. Then, based on
the insightful mechanisms of the multi-population evolution, cascade correlation,
architecture, and strategy, as well as optimization methods, the LECO model is developed.
The architecture of the LECO endorses hybridization of different optimization techniques.
Different combinations of GA and PSO in LECO are studied to investigate the most
appropriate combination of evolutionary mechanism for LECO. A series of empirical
studies comprising benchmark and real-world problems is employed to assess the capability,
flexibility, and effectiveness of LECO to handle different kinds of optimization problems as
well as to be integrated with other heuristic techniques. Besides the datasets given in
benchmark and real-world problems, hypothetical data is also included to investigate the
performance of LECO towards larger scale problems. The experimental results
demonstrate that GA-PSO LECO is able to optimize combinatorial multi-decision
scheduling problem, multi-resolution parameter optimization as well as multi-objective
Pareto optimization. In addition, the LECO structure that allows particular layer to be
easily analyzed and evaluated promotes interactive optimization whereby human
xix
intervention can be applied on layers for feature extraction. In all, the cascade layered
encoding structure that is able to show inheritance of information, separating representation
into layers, enhancing balance global-local search, and narrowing down the search space
makes the LECO model a flexible, generic, and powerful tool for optimization problems.
Description
Keywords
The Development And Application Of Evolutionary Computation-Based , Layered Encoding Cascade Optimization Model