Research Of Web System Parameters Optimization Based On BPSO Algorithm

The Internet affects our life profoundly. A challenge that we should make full use of theWeb system performance to provide more services is put forward by growing user. Adjustingthe parameters of the Web system can significantly improve the performance of system. Theparameters of the Web system is numerous. So adjusting parameters is difficult and the staffwho does that need rich experience.In this paper, binary particle swarm optimization(BPSO) algorithm and its improvedalgorithm are proposed to optimize Web system parameters by analyzing the self-built Websystem and integrating various aspects of condition. Finding the optimal configurationparameters combination is the purpose. Then we get the best Web system performance. Thispaper mainly includes following content:(1) The deployment of experimental platform. After analysis, the Web system that iscomposed of Apache, MySQL and PHP is selected to study. ApacheBench is selected to beperformance test tool. Control system that is compiled by Python connects Web system andperformance test tools, and the deployment of experimental platform is completed.(2) The realization of the Web system parameters optimization by using binary particleswarm optimization(BPSO) algorithm. We analyze the basic principle, working process andthe factors influencing the performance of binary particle swarm optimization(BPSO)algorithm and apply it to the Web system parameter optimization problem. We encode themain adjustable configuration parameters in Web system, use computer random to initializethe population and obtain the global optimal solution through the iterative optimization ofbinary particle swarm optimization(BPSO) algorithm. Then the experiment results andalgorithm performance analysis are given.(3) To improve binary particle swarm optimization(BPSO) algorithm, then use theimproved algorithm to realize the Web system parameters optimization. We find some problem of binary particle swarm optimization(BPSO) algorithm through analysis. It is easyto premature. Its diversity of particles reduces too quickly in the early stage and local searchability is weak in the later stage. Complement each other, we introduce dissipation operationto enhance the global search ability in the early stage and hill-climbing(HC) algorithm toenhance the local search ability in the later stage. Then those form a new hybrid algorithm.We use new hybrid algorithm to reliaze the Web system parameter optimization and obtainthe global optimal solution. Then the experiment results and algorithm performance analysisare given.In this paper, we deploy the experimental platform to test Web system performanceunder the limited resources and load. We run the algorithm program to find the optimal ornear optimal system configuration parameters and provide the experimental results andanalysis. The validity of the optimization algorithm is verified. The optimization algorithmhas practical application value.