An improved reinforcement learning approach to solve flow job scheduling problems

The fow shop scheduling problem is an important scheduling problems with a large number of practical applications. For the high performance computing applications, flow shop scheduling directly affects resource utilization and power dissipation.In this paper, we propose a novel algorithm TS_Qlearning algorithm that combines the tabu search(TS) method and the Qlearning algorithm to minimize the idle time. We also calculate the makespan value,which is used to comprehensively evaluate the quality of the algorithm. The comparative analysis of the results proves the superiority of the TS_Qlearning algorithm.