運用菁英跳脫策略改良人工蜂群演算法

Improving Artificial Bee Colony Algorithm with Elite Escaping Strategy

楊清淵(Ching-Yuan Yang)

指導教授 ： 李維平

中原大學/商學院/資訊管理研究所/碩士(2012年)

https://doi.org/10.6840/cycu201200377

摘要

人工蜂群演算法是近年來新穎的演算法，已經被證實優於基因演算法、粒子群演算法、差分演算法等，但其在精準度、收斂速度仍有可提升的空間，也同樣具有演算法易陷入區域最佳解的問題。
本研究提出改良工蜂和觀察蜂的搜尋公式，利用最佳解向量以及個體最佳解來引導搜尋的方向，幫助搜尋解以及收斂的速度，在偵查蜂的部分，加入了全體最佳解和個體最佳解向量來產生，並使用判斷閥值來避免陷入區域最佳解的問題，又能有效提升求解精度。
本研究則主要提出四個實驗，分別是自己方法比較、相關研究比較、雜訊實驗、非0最佳解，可以驗證出在不同實驗環境下，仍有優異的效能，未來更可以再嘗試結合其他演算法。

關鍵字

全域最佳解 ； 個體最佳解 ； 菁英跳脫策略 ； 人工蜂群演算法

並列摘要

The artificial bee colony algorithm ,a novel algorithm in recent years, has been shown to be superior to Genetic algorithms, Particle Swarm Optimization algorithms, and Differential Evolution algorithm. However,there is still room for its improvement in accuracy and convergence rate; in addition, it also involves the issue in terms of the problem of the algorithm,which is easy to fall into the local optimal solution.
This study proposes a search formula for improved employed bees and onlookers bee, by the use of optimal solution vector as well as individual optimal solution to guide the search direction, and to help search for solutions and the speed of convergence. In the scout bee,adding global optimal solutions and individual optimal solution vector to generate and using it to determine the threshold to avoid the problem of the local optimal solution,can effectively enhance the solution accuracy.
This study consists in three experiments, including comparison with related researches, experimental noise, and non-0 the optimal solution. The main objective of the three experiments verifies the outstanding performance in different experimental environments. Different combination with other algorithms can also be examined in the future.

並列關鍵字

Artificial Bee Colony algorithm ； Global optimal solutions ； Elite Escape Strategy ； Individual optimal solution

參考文獻

林豐澤（2005）。演化式計算下篇:基因演算法以及三種應用實例。智慧科技與應用統計學報，3(1)，29-56。

郭定（2009，A）。基因演算法中不同選擇策略的替代性與互補性。科學與工程技術期刊，5（2），25-34。

葉進儀、林彣珊、朱慶餘（2007）。應用平行基因演算法改善護理人員排班品質。品質學報，14（3），337-350。

Aderhold, A., Diwold, K., Scheidler, A., & Middendorf, M. (2010). Artificial bee colony optimization: A new selection scheme and its performance. Computational Intelligence,284(2),283-294.

Alatas, B. (2010). Chaotic bee colony algorithms for global numerical optimization. Expert Systems with Applications, 37(8), 5682-5687.