Reactive Hybrid Particle Swarm Optimization Based Job-shop Scheduling Problems Considering Energy Management

Improved Parallel Reactive Hybrid Particle Swarm Optimization using Improved Neighborhood Schedule Generation Method for the Integrated Framework of Optimal Production Scheduling and Operational Planning of an Energy Plant in a Factory

“…For 10job 10-machine benchmark problem, the optimal makespan of 943 was obtained with HPSO 19 only at TL length of 40, while RHSPO offered the optimal makespan at any initial TL length according to the literature. 21 In this paper, too, the optimal makespan was obtained at any initial TL length for all proposed methods of improved parallel RHPSO. Figure 8 illustrates how average calculation time depends on the number of processes in improved parallel RHPSO (proposed method 3).…”

“…19 The authors used reactive HPSO (RHPSO) in which TS module of HPSO is replaced with RTS (conventional method 2), and confirmed better quality of solutions as compared to the conventional method 1. 21 However, with the conventional method 2, only order of adjacent job pairs was changed in a CB, which posed a problem in neighborhood schedule generation. Thus, there still was room for F I G U R E 2 Integrated framework of primary problem and subproblem improvement of solution quality to secure diversity in neighborhood schedule generation using CB s.…”

“…As regards details of the entire RHPSO algorithm, please refer to the literature. 21 Step 1. A storage list is created using a hash function, and initial TL length is set.…”

“…Job-shop-type production scheduling has shifted to highmix, low-volume production, and various schedules must be obtained within a short time; that is, fast scheduling is needed. However, RHPSO 21 is problematic in that multipoint search is sequentially processed that results in a long calculation time. In addition, time required for RTS in local search at Step 7 of the improved RHPSO algorithm described in Section 3.1 accounts for about 99.92% of the entire calculation time; that is, cutting the local search time would greatly contribute to the overall calculation speed.…”

“…In this paper, 10-job 10-machine problem (abz6 26 ) was used as a benchmark, and the proposed improved parallel RHPSO and conventional method 2 21 were applied to JSP with regard to minimization of secondary energy costs; the simulations were intended to verify improvement of solution quality and acceleration by parallel processing using OpenMP. 27 In addition, parameters of the proposed methods were varied assuming application to high-mix, low-volume production.…”