An Efficient Local Search Operator Using Orthogonal Experimental Design for Genetic Algorithm Optimization

• Main Authors: Yan-Fan Chen, 陳彥帆
• Other Authors: Shinn-Ying Ho
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/wn9p4n

碩士 === 逢甲大學 === 資訊工程所 === 91 === In this thesis, an orthogonal experimental design (OED) based local search operator is proposed to enhance the search ability of genetic algorithms. Traditional local search operators use hill-climbing strategies to generate and test a number of candidate solutions in the neighborhood of a current solution, and then replace the current solution with the best candidate one. Recently, a genetic algorithm with a simplex method as the local search operator is proposed for function optimization. Besides, various improved mutation operators are used in evolutionary programming to solve the function optimization problem successfully. The proposed local search operator embedded in genetic algorithms systematically samples a small number of candidate solutions and then reasons a potentially good approximation to the best solution in the neighborhood of the current solution. Experimental results show the effectiveness of the hybrid genetic algorithm with the proposed local search operator using thirteen benchmarks, compared with some existing evolutionary programming and evolutionary parallel local search algorithms.