本論文之目的係為增加群體內粒子的搜尋能力和效率,而藉由模糊理論的觀念來調適粒子群演算法的加速度參數,這種策略主要是希望粒子群能持續的拓展搜尋範圍和找尋新的最佳解。本論文有兩種優點可以描述,一種就是可以和其它不同型式粒子群演算法的加速度參數做結合增加他們的搜尋性能,另一種就是經由模糊理論訂出三條模糊規則做加速度參數調整,達到全域搜尋能力。另外為了分析演算法於不同領域和適用性問題,透過16種標準函數的測試與5種不同的粒子群算法做比較。最後,經由模擬的結果顯示,本方法的確可以有效地改善原始PSO的性能,並且對於大多數的標準測試函數而言,均有優越的表現。In this thesis, in order to enhance each variable particle’s searching ability and efficiency, a fuzzy logic control is implemented to adapt the acceleration parameters of particle swarm optimization algorithm (PSO). The important condition of fully utilizing the particle swarm optimization algorithm is to keep advance between extensive searching and exploring global optimal. This method has two advantages. One is that it is flexible to integrate with other PSO techniques to enhance the searching performance further. The other is that it is only used three simple fuzzy inference rules to adaptively adjust the acceleration parameters of the standard PSO and results in certain improved searching ability and efficiency. In addition, the simulation is tested by using 16 benchmark functions. The results show that our proposed methods can efficiently improve the performance of original PSO and outperform the five compared PSO algorithms for most of benchmark functions.