Abstract: Aiming at the disadvantages that the Particle Swarm Optimization（PSO） algorithm easily fall into local optimization and low search accuracy for solving high-dimensional and complex optimization problems, this paper proposes a hybrid particle swarm optimization algorithm, named CMCPSO, which borrows the idea of colony mountain climbing from Shuffled Frog Leaping Algorithm（SFLA）, and the global convergence of CMCPSO is also proved. Solving four typical high-dimensional continuous optimization functions shows that CMCPSO not only maintains the fast convergence of PSO algorithm, but also absorbs the advantages of SFLA algorithm for local refined search and maintaining the diversity of the population, it also has a good global convergence property.

Key words: Particle Swarm Optimization（PSO）, shuffled frog leaping algorithm, colony mountain climbing strategy, global convergence, function optimization

摘要： 针对粒子群优化算法（PSO）在求解高维复杂优化问题时存在搜索精度不高和易陷入局部最优解的缺陷，借鉴混合蛙跳算法（SFLA）的群体爬山思想，提出一种基于群体爬山策略的混合粒子群优化算法（CMCPSO），并证明了CMCPSO算法的全局收敛性。对四个典型高维连续优化函数的求解表明，该算法不仅保持了PSO算法的快速收敛能力，而且吸收了SFLA算法局部精细搜索和保持种群多样性的优点，具有良好的全局收敛性。

关键词: 粒子群优化算法, 混合蛙跳算法, 群体爬山策略, 全局收敛性, 函数优化