基於多樣性慣性權重策略粒子群最佳化演算法之研究

A Study on Particle Swarm Optimization Algorithm by Diversity-Based Inertia Weight Strategy

廖雲枝(Yun-Chih Liao)

指導教授 ： 李維平

中原大學/商學院/資訊管理學系/碩士(2007年)

https://doi.org/10.6840/cycu200700284

摘要

粒子群最佳化演算法(PSO)源自於對鳥群、魚群和人類社會行為的觀察而提出的一種最佳化技術，其主要特點為簡單且容易實現，PSO同時具備進化演算法和群體智慧的特徵，它透過群體中粒子的合作與競爭而產生群體智慧，吸引著所有粒子往最佳解的方向搜索，而且已經在數值最佳化等問題中表現出良好的求解能力，但是粒子群最佳化演算法仍存在容易陷入區域最佳解及搜索精度不高等缺點。
粒子於演化搜尋過程中，隨著粒子逐漸的收斂，群體多樣性也會逐漸降低，導致粒子處於停滯現象，因此維持群體的多樣性，可藉由提高粒子的活性，避免粒子早熟和陷入區域解。針對標準PSO之慣性權重係數係採固定或線性遞減的方式，無法有效的解決粒子陷入區域解的問題及可能出現的停滯現象，本研究引入粒子多樣性的機制和多樣性的活化策略，予以動態調整慣性權重參數及粒子的活性，促使粒子收斂到全域最佳解，以改善粒子群最佳化演算法的開發和探索能力，並提高其收斂速度及精度。

關鍵字

多樣性 ； 粒子群最佳化演算法 ； 群體智慧

並列摘要

Particle swarm optimization algorithm (PSO) is motivated from the simulation of simplified social behavior of bird flocking, fish schooling and human. The PSO is simple and easy to implement on many optimization problems. And it is demonstrated that it is an efficient way to solve the optimization question. However, it doesn’t overcome the problems of avoiding premature convergence and escaping local optima.
On multi-modal test problems the PSO tends to suffer from premature convergence and a decrease of diversity in search space. That leads to fitness stagnation of swarm. Therefore to keep high diversity is crucial for preventing premature convergence and trapping in local solution. In the classical PSO, a constant or linearly decreasing inertia weight is used for solving the optimization problem, but it is unable to solve the phenomenon of stagnation.
In this research, a diversity-based inertia weight strategy and the activation of swarm in the PSO are proposed. The results show that a diversity-based inertia weight strategy for PSO improves exploitation and exploration ability, but still keeps a rapid convergence and still fine precision.

並列關鍵字

particle swarm optimization algorithm ； diversity ； swarm intelligence

參考文獻

[11] Arumugam, M. Senthil and Rao, M. V. C. “On the performance of the particle swarm optimization algorithm with various inertia weight variants for computing optimal control of a class of hybrid systems,” Discrete Dynamics in Nature and Society 2006, Article ID 79295, pp. 1-17, 2006.

[9] Angeline, P. J., “Evolutionary optimization versus particle swarm optimization: philosophy and performance differences,” Evolutionary Programming VII: Proceedings of the Seventh Annual Conference on Evolutionary Programming, 1998.

[10] Angeline, P. J., “Using selection to improve particle swarm optimization,” Proceedings of the IEEE Congress on Evolutionary Computation (CEC 1998), Anchorage, Alaska, USA. 1998.

[13] Bonabeau, E., Dorigo, M., and Theraulaz, G. “Swarm Intelligence: From Natural to Artificial Systems,” New York: Oxford University Press, 1999.

[15] Clerc, M., and Kennedy, J. “The particle swarm–explosion, stability, and convergence in a multidimensional complex space,” IEEE Transactions on Evolutionary Computation, vol. 6, no. 1, pp. 58-73, 2002.