- 學位論文
基因演算法與粒子群演算法之比較-以派課問題為例
Comparison of Genetic Algorithm and Particle Swarm Optimization for the Course Assignment Problem
摘要
基因演算法(Genetic Algorithm, GA)是一種模擬生物界自然選擇與進化機制所發展出之高度平行、隨機、自適應的最佳化搜尋方法，其透過演化的機制隨機地逐步淘汰較差的個體，保留較好的個體。而粒子群演算法(Particle Swarm Optimization, PSO)模擬生物之群體行為進行分群，是近年來相當受到重視之新興智能搜尋方法，其特徵在於演算法易於實現，快速和優於其他智能演算法的全域搜尋能力。粒子群演算法產生群體與計算適應值的步驟與基因演算法相當類似，但兩者後期演算步驟與執行效能皆不盡相同。本研究期望透過大專院校之派課問題(Course Assignment Problem)，比較粒子群演算法與基因演算法其執行成本的差異性。
並列摘要
Particle Swarm Optimization (PSO) is a popular optimize search method that inspire by the swarming of biological populations in recent years. PSO is similar to the Genetic Algorithm (GA) in the sense that both are population-based search method. Since the two methods are supposed to find a solution to given objective function but employ different procedure and computational effort, it’s appropriate to compare their implementation. The disadvantage of the GA is its expensive computational cost. This paper attempts to examine that PSO has the same efficiency as the GA but with better computational efficiency. The problem area of the PSO and GA chosen is Course Assignment Problem (as used in the university).