Abstract: With the dramatic increase in the amount of network users, Web server is widely used in large software systems and usually needs to take a lot of effort to configure multiple parameters related with performance before running. The process of manual configuration parameters is too cumbersome and requires professional knowledge and experience. To obtain the reasonable web system configuration parameters conveniently and quickly, an optimization algorithm of web system based on hybrid binary particle swarm optimization algorithm is presented. The algorithm which joins the experience factor, hill-climbing algorithm and the linearly decreasing inertia weight can automatically find the optimal configuration parameter of Web system. The algorithm can solve the problems of traditional Binary Particle Swarm Optimization（BPSO） algorithm which has low efficiency and easy to fall into local optimal solution. Experimental results show that the algorithm has high efficiency, can jump out of local optimal solution, and can get better global optimal solution.

Key words: optimization of Web system, hybrid binary particle swarm optimization, experience factor, mountain climbing algorithm, linearly decreasing inertia weight

摘要： 随着网络用户量的急剧增加，Web服务器被广泛应用于大型软件系统中，系统在运行前通常需要配置与性能相关的多个参数。人工配置参数的过程太繁琐且需要专业知识与经验，为了更便捷、更快速获取合理的系统配置参数，提出了一种基于混合二进制粒子群的Web系统优化算法。该算法加入了经验因子、爬山算法、线性递减惯性权重，对Web系统自动迭代寻找最优配置参数，解决了传统二进制粒子群算法寻优效率低、容易陷入局部最优解等问题。实验结果表明，该算法寻优效率高，能跳出局部最优解，可以获得效果更好的全局最优解。

关键词: Web系统优化, 混合二进制粒子群优化, 经验因子, 爬山算法, 线性递减惯性权重