Abstract: Genetic algorithm (GA) and particle swarm optimization (PSO) have been broadly used in many fields, such as function optimization, production scheduling, machine learning and data mining, etc. This paper makes comprehensive and systematic comparisons of GA and PSO on dealing with a series of benchmark problems, analyzes their performance on solving unimodal and multimodal functions, and further tests the robustness of two algorithms for investigating the influences of parameters to the performance of the algorithms. This paper finally concludes the characteristics of the two algorithms and discusses their improvement strategies. The goal of this paper is to provide technical guidance for the selection of algorithms in engineering applications.

Key words: genetic algorithm, particle swarm optimization, unimodal, multimodal, performance comparison

摘要： 遗传算法与粒子群优化算法作为经典的进化计算方法已经被广泛地应用于函数优化、生产调度、机器学习和数据挖掘等领域。对这两种经典算法在求解不同问题时的性能进行了系统的对比和分析，比较了两种算法在求解单峰和多峰问题上的性能差异。进一步对算法的健壮性进行了测试，分析了算法运行过程中参数对算法性能的影响。最终总结出两种算法的性能特点，并讨论了算法的改进策略，旨在为工程应用中的算法选择提供技术参考。

关键词: 遗传算法, 粒子群优化算法, 单峰, 多峰, 性能对比