本文旨在提出廣義性最佳化演算法探討與改良，其中廣義的最佳化演算法為最受重視，重要原因是適用範圍層面的問題，基因最佳化演算法、粒子群最佳化演算法、模擬退火法…等最常被使用，在規律的適應函式找極值上，每個方法各有優缺點，而優缺點取決於是否適應值符合標準和達到目標花費時間上的多寡，在不同的適應函式也各有不同的取向，基於以上的需求，本文也試著尋找更優於一般廣義性最佳化演算法的演算方式。 本論文嘗試以高度指向性配合強烈具跳出局部能力之方式，以此來達到搜索更迅速之目的，加上精簡演算法流程與使執行者使用上更方便設計的目標來完成此演算法，文末，吾人亦提出模擬以驗證本文改良法之可行性。 This study aims to explore the problems of strengthen-reverse and reduction for particle swarm optimization. Genetic algorithm, particle swarm optimization and simulated annealing are widely used to search for the global optimal solutions of fitness functions. The present work tries to make some improvements and to reduce the consuming time of the generalized optimization algorithms. Whether the generalized optimization algorithms are good or bad usually depends on the fitness function value. This paper tried to use high pointing-behavior to make the speed of seeking out the global optimum being higher. But we need to increase the chance of escaping from the local optimal solutions. Final the new algorithms are as simplified as possible and that the user will apply these algorithms more easy than others. In addition, the simulation is given to verify the feasibility of the present method.
