针对目前研究相对薄弱的群体智能优化算法的性能对比问题,搭建数字图像为生命栖息环境的群体智能优化算法的性能对比平台,提出基于最优个体变化的收敛关联度和收敛面积的新型性能评价指标,并具体进行了遗传算法、粒子群算法、人工鱼群算法、细菌觅食算法等多种群体智能优化算法的性能比较与测试.实验结果显示,所提出的评价平台和性能评价指标能够合理有效地对比不同搜索机制下智能群体的寻优能力.
Aiming at the performance comparison of swarm intelligence optimization algorithms that lacks qualified research findings, we constructed a platform for comparing the performance of the algorithms. Then, we proposed the novel performance evaluation criteria for convergence relational degree and the convergence area based on the changes of the best individual. Specifically, we compared and tested the performances of several swarm intelligence optimization algorithms, such as the genetic algorithm (GA), particle swarm optimization (PSO) algorithm, artificial fish swarm (AFS) algorithm, bacterial foraging (BF) algorithm and artificial bee colony (ABC) algorithm. Experimental results showed that the platform and criteria of performance evaluation proposed in this paper can be effectively used to compare the capability of optimization search under different mechanisms.