An effective TPA-based algorithm for job-shop scheduling problem

The job-shop scheduling problem (JSSP) is one of the most general and difficult of all traditional scheduling problems. Many different approaches have been applied to JSSP and a rich harvest has been obtained. However, some JSSP, even with moderate size, cannot be solved to guarantee optimality. In this paper, a computationally effective team process algorithm (TPA) for solving the minimum makespan problem of job-shop scheduling is used. In the TPA system, the team numbers are divided into the elite and plain groups. The manipulations of learning and exploring are properly defined, and the member renewal rules are reasonably established. It makes the algorithm possesses the potential of global, local and directional search. Numerical results verify that the modified TPA has the properties of simple implementation, high success rate for global optimization, fast convergence, and better than the modified particle swarm optimization (PSO) with GA and GA alone.