A Combined Algorithm for Solving Job Shop Scheduling Problems

In this paper, a combined algorithm consisting of two stages of approaches is proposed for solving a Job Shop Scheduling Problem (JSSP). The first stage of approach called Semi-active Scheduling Approach (SSA) is proposed for obtaining a local optimal solution in a short time by using the improved local search algorithm and the neighborhood search technique proposed in our earlier paper. The second stage of approach called Active Scheduling Approach (ASA) is proposed for improving the solution obtained from SSA and preventing the solution from trapping in the local minimum by reducing the idle time in the processes and making good use of the resources. Both of the approaches focus their improvement efforts on reducing production expense. The proposed algorithm is applied to solving JSSPs. A large number of computational experiments show that the combined algorithm can overcome the disadvantages of the respective approaches, SSA and ASA, and obtain a better solution as compared with single use of such approaches for solving the complicated JSSPs. In addition, the proposed algorithm is compared with Genetic Algorithm (GA), and the result shows that in the case of limited total search points, the proposed algorithm can converge to a suboptimal solution faster than GA.