Characterizing the performance of adaptive swarm-based simulated annealing algorithm

Date of Publication

2009

Document Type

Bachelor's Thesis

Degree Name

Bachelor of Science in Chemical Engineering

College

Gokongwei College of Engineering

Department/Unit

Chemical Engineering

Thesis Adviser

Raymond Girard Roca Tan

Defense Panel Member

Cesar A. Llorente
Florinda Tiangco Bacani

Abstract/Summary

Swarm-based simulated annealing (SBSA) is a hybrid algorithm based from simulated annealing with the incorporation of swarm intelligence to resolve the problem of slow convergence in obtaining the optimal solutions. This study focuses on the characterization of algorithm by adjusting the parameters which affect its performance in terms of accuracy, consistency, and computational efficiency. The parameters which affect the performance of the algorithm were determined by solving eleven (11) test problems with different number of variables, constraints, and mathematical operators. The parameters that were varied were the swarm size, outer loop iteration, adaptive cooling coefficient, and step size, which were further characterized by assigning low and high level values in a two-level factorial experimental design. A configuration which has a higher parameter value of swarm size and number of iterations and a lower parameter value of step size and adaptive cooling coefficient was found to produce near-optimal and consistent solutions in test problems with small number of variables, constraints, and mathematical operators. The solutions of test problems were obtained at the low level value of outer loop iteration. However, increasing the number of outer loop iterations can increase the chance of finding the global optima. Increasing the penalty function and swarm size can increase the probability of exhibiting near-optimal solution for large-scale test problems. SBSA program is not recommended to be used for large-scale test problems because it produced inaccurate and inconsistent data based from the results of the study.

Abstract Format

html

Language

English

Format

Print

Accession Number

TU15947

Shelf Location

Archives, The Learning Commons, 12F, Henry Sy Sr. Hall

Physical Description

113 leaves : ill. ; 28 cm.

Keywords

Swarm intelligence; Simulated annealing (Mathematics)

This document is currently not available here.

Share

COinS