Abstract

In this article, an improved Gaussian distribution based quantum-behaved particle swarm optimization (IG-QPSO) algorithm is proposed to solve engineering shape design problems with multiple constraints. In this algorithm, the Gaussian distribution is employed to generate the sequence of random numbers in the QPSO algorithm. By decreasing the variance of the Gaussian distribution linearly, the algorithm is able not only to maintain its global search ability during the early search stages, but can also obtain gradually enhanced local search ability in the later search stages. Additionally, a weighted mean best position in the IG-QPSO is employed to achieve a good balance between local search and global search. The proposed algorithm and some other well-known PSO variants are tested on ten standard benchmark functions and six well-studied engineering shape design problems. Experimental results show that the IG-QPSO algorithm can optimize these problems effectively in terms of precision and robustness compared to its competitors.

摘要

在本文中，提出了一种改进的基于高斯分布的量子行为粒子群优化（IG-QPSO）算法来解决具有多个约束的工程形状设计问题。在该算法中，采用高斯分布来生成 QPSO 算法中的随机数序列。通过线性减小高斯分布的方差，该算法不仅能够在早期搜索阶段保持其全局搜索能力，而且在后期搜索阶段也可以获得逐渐增强的局部搜索能力。此外，在 IG-QPSO 中采用加权平均最佳位置来实现局部搜索和全局搜索之间的良好平衡。所提出的算法和其他一些著名的 PSO 变体在十个标准基准函数和六个经过充分研究的工程形状设计问题上进行了测试。实验结果表明，与竞争对手相比，IG-QPSO 算法可以在精度和鲁棒性方面有效地优化这些问题。