Improve performance of a material control system using parallel processing with particle swarm optimization

A timeout problem occurs when multiple jobs with increasing data are concurrently competing for limited computing resources. Adopting a particle swarm optimization (PSO) method, this work proposes a parallel processing scheme to improve the efficiency of job execution based on available virtual machines. In this scheme, a map-reduce like architecture with cache preloading is presented to execute the jobs according to the PSO scheduled execution plan. The results of a material control system case show that the mean time for executing reporting jobs has been reduced by using cache preloading. Especially, the constraint of the time to job completion can be maintained by using this proposed scheme by adding the number of parallel virtual machines in the map-reduce like architecture to answer increasing of the processing data.