- 期刊
以多樣性活化策略改良粒子群演算法求解成效
Improving the Performance of Particle Swarm Optimization Algorithms with Diversity-Based Activation Strategy
摘要
目前對於粒子群最佳化演算法(PSO)的改良研究均針對演算法參數或其他改良方式進行改善以避免粒子提早收斂，進而維持演算法逃脫區域解的能力，但對於粒子群陷入局部最佳解的處理方式卻較少被討論。因此本研究嘗試使用以多樣性為基礎的活化策略對於PSO進行改良，透過實驗結果證明本研究有三項優點，首先，加入本研究的活化策略以後，演算法在求解精度上確有大幅的改善，其中以非線性權值遞減策略 (EDIW)的PSO最為顯著。另外，由於現有PSO的實務應用研究仍多以早期的慣性權值線性遞減策略為主，因此本研究可提供後續採用早期慣性權值線性遞減策略的PSO進行應用研究的參考。最後對於後續欲採用非線性權值遞減策略(EDIW)的實務應用研究而言，本研究亦可提供作為更佳的參考策略。
並列摘要
According to researches of Particle Swarm Optimization algorithm (PSO), strategies of preventing premature convergence and trapping in local solution were more than the particle really trapped in local solution. This article adopted the concept of diversity-based activation strategy to improve Particle Swarm Optimization algorithm. The experimental results show three advantages for the activation strategies. First, it's workable for the algorithm and the performance was superior then the original algorithm. Second, it provided useful information for the applications which's algorithm with linearly decreasing inertia weight. Third, it can provide great performance to the applications which's algorithm with exponential decreasing inertia weight.
延伸閱讀
- Chang, C. C. (2015). 以詢問式學習機制改良粒子群演算法應用於動態環境之研究 [master's thesis, National Taiwan University]. Airiti Library. https://doi.org/10.6342/NTU.2015.10555
- 陳志忠（2011）。含突變機制的粒子群演算法於多目標工程最佳化〔碩士論文，淡江大學〕。華藝線上圖書館。https://doi.org/10.6846/TKU.2011.00991
- 莊佾叡（2009）。改良型粒子群演算法於上行多輸入多輸出分頻多工系統多用戶偵測之應用研究〔碩士論文，國立臺北科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0006-2107200913581500
- 廖雲枝（2007）。基於多樣性慣性權重策略粒子群最佳化演算法之研究〔碩士論文，中原大學〕。華藝線上圖書館。https://doi.org/10.6840/cycu200700284
- Chen, X. H. (2007). A Collaborative and Adaptive Approach to Particle Swarm Optimization [master's thesis, Chung Yuan Christian University]. Airiti Library. https://doi.org/10.6840/cycu200700542