Testing multiple threads Tabu search by solving scheduling problems

Abstract

This study addresses the efficiency of multiple threads Tabu search (TS) in solving scheduling problems. Nowadays, most desktop personal computers equip with multicore CPU. It is possible to achieve parallel searching strategy on a desktop computer. A problem of scheduling, which minimizes the total tardiness of a set of jobs to be scheduled on parallel identical machines, is presented as an example in this study. The parallel program uses Open Multi-Processing (OpenMP) library under Microsoft Visual Studio 2010 environment. An experiment of different number of jobs is executed by the parallel Tabu search program with different number of threads, and the results of single thread and multithreads are discussed. In sum, the more threads used for searching process, the lager instances to get a better solution.