Integration of process planning and job shop scheduling with stochastic processing time

In a conventional approach, process planning and scheduling are two separate tasks and perform sequentially. This is while without the integration of process planning and scheduling, a true computer-integrated manufacturing system may not be effectively realized. Although there are several scientific manuscripts which address some approaches for integration of process planning and scheduling in the recent years, the focus of these researches was on deterministic constraints of jobs. In this paper, we consider stochastic parameters and present a new approach to adjust to the real-world industry situations. In this way, the CAPP system generates all the possible process plans at first, and then four near-optimal process plans are selected via Dijkstra algorithm, and ten scenarios are generated with Monte Carlo sampling method. A mathematical model was solved within reasonable time with a hybrid algorithm consisting of Simulated Annealing and Tabu Search. To evaluate the proposed algorithm, four problems were generated and solved with the proposed algorithm in the deterministic and stochastic manners, which indicates that stochastic results are more robust than those of deterministic in different situations. Then, the same experiments were solved taking advantage of Lingo to evaluate the running time, which shows that the hybrid algorithm exhibits high performance in large-scale problems, whereas the running time of Lingo was increased exponentially. As a result, the proposed algorithm generates solutions in more acceptable time than Lingo, especially for large-scale problems.