啟發式鯨魚演算法於結構最佳化設計之研究

Optimum Design of Structures by a Meta-heuristic Whale Optimization Algorithm

江舫萱(Fang-Hsuan Chiang)

指導教授 ： 張永康

淡江大學/工學院/航空太空工程學系碩士班/碩士(2018年)

https://doi.org/10.6846/TKU.2018.00343

摘要

本論文應用鯨魚演算法於結構最佳化設計中。鯨魚演算法是一種模仿自然界座頭鯨覓食行為進行搜尋最佳解之方法，經由將座頭鯨捕食獵物之過程透過數學轉換為一仿生最佳化演算法。鯨魚演算法可分為三種搜尋方式，分為探勘階段以及開發階段中的收縮環繞和螺旋更新位置。鯨魚演算法的優點為架構簡單且具有較少的參數需要設定，並以隨機的方式進行搜尋，因此可以做大範圍之全域搜尋。本研究提出改良式鯨魚演算法，採用以亂數產生之步伐係數，使演算法進入區域搜索時能有彈性的搜索空間。為了避免產生無效率的搜索，若目前位置就是當前最佳位置程式即自動產生新解，如此能有效的避免落入區域最佳解，並且可以迅速的搜尋全域最佳解。
　　範例中將結構最佳化問題轉為數學函數，再利用鯨魚演算法對結構系統執行最佳化設計。由數值分析範例之結果，發現應用鯨魚演算法於結構最佳化設計上可得到不錯的結果。

關鍵字

鯨魚最佳化演算法 ； 最佳化設計

並列摘要

The Whale Optimization Algorithm (WOA) was applied to the optimum design of structures in this study. The WOA algorithm is swarm intelligence based optimization technique inspired by the intelligent foraging behavior of humpback whales. There are three search methods in WOA, which are exploration phase, shrinking circle mechanism and spiral updating position. The advantages of WOA are simple concept, wide searching range and less control parameters needed. WOA can search in a random manner, so the global search can be performed efficiently with the searching techniques. This study proposes an enhanced whale optimization algorithm, which uses the different influence of the coefficient, so that the algorithm have a flexible searching space when entering the local search manner. In order to avoid the inefficiency search, a new position will be generated automatically if the current position is the current best position. This scheme can help searching the global optimum quickly and preventing the design fall into local optimum effectively.
　　The optimization problem can be transformed into a mathematical function. Minimum weight design will be discussed in numerical examples. The results of WOA are better than other references in the numerical examples.

並列關鍵字

Whale Optimization Algorithm ； Optimum Design

參考文獻

[1] Holland, J. H., “Genetic programming,” Scientific American, Vol. 267, pp.66-72, 1992.

[2] Rechenberg, I., “Evolutionsstrategien,” Vol. 8, Springer Berlin Heidelberg, pp.83-114, 1978.

[3] Kirkpatrick, S., Gelatt, C. D. and Vecchi, M. P. , “Optimization by Simulated Annealing,” Science, New Series, Vol. 220, No. 4598, pp.671-680, 1983.

[4] Shen, J. and Li, J., “Local Convergence Analysis of Greedy Light Ray Optimization Algorithm,” Journal of Harbin Engineering University (Science Edition), Vol. 50, No. 2 , pp.208-212, 2012.

[5] Eberhart, R.C. and Kennedy, J., “Particle Swarm Optimization,” IEEE International conference on neural networks, Vol. 4, pp.1942- 1948, 1995.