Integrating Tabu Search into the Survivor Strategy of Genetic Algorithm

• Main Authors: Cheng-Feng Ko, 柯政鋒
• Other Authors: Chuan-Kang Ting
• Format: Others
• Language: en_US
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/09807277415209858676

碩士 === 國立中正大學 === 資訊工程所 === 95 === Genetic algorithm (GA) is very effective to solve optimization problems. However, GA suffers from the serious problem of premature convergence. Premature convergence will trap GA into local optima. To avoid this problem, some studies solve it by maintaining population diversity. Tabu genetic algorithm (TGA) applies a new mating strategy to maintain diversity through preventing inbreeding chromosomes. However, the selection process in TGA can be very time-consuming. In this thesis, we propose a novel hybrid method, TGA2, to prevent premature convergence. TGA2 integrates strategy of tabu search into the “survival” procedure of GA. The performance of the proposed TGA2 is evaluated and compared to GA and TGA on six numerical optimization problems and four combinatorial optimization problems. TGA2 outperform the solution quality than GA on these problems. In addition, TGA2 using relaxed aspiration criterion accelerates convergence speed. Experimental results validate that the relaxed aspiration criterion helps to speed up the convergence of TGA and TGA2.