An evolutionary approach for solving the multi-objective job-shop scheduling problem

In this chapter, we present an evolutionary approach for solving the multi-objective Job-Shop Scheduling Problem (JSSP) using the Jumping Genes Genetic Algorithm (JGGA). The jumping gene operations introduced in JGGA enable the local search process to exploit scheduling solutions around chromosomes, while the conventional genetic operators globally explore solutions from the population. During recent decades, various evolutionary approaches have been tried in efforts to solve JSSP, but most of them have been limited to a single objective, which is not suitable for real-world, multiple objective scheduling problems. The proposed JGGA-based scheduling algorithm heuristically searches for nearoptimal schedules that optimize multiple criteria simultaneously. Experimental results using various benchmark test problems demonstrate that our proposed approach can search for the near-optimal and nondominated solutions by optimizing the makespan and mean flow time. The proposed JGGA based approach is compared with another well established multi-objective evolutionary algorithm (MOEA) based JSSP approach and much better performance of the proposed approach is observed. Simulation results also reveal that this approach can find a diverse set of scheduling solutions that provide a wide range of choice for the decision makers. © Springer-Verlag Berlin Heidelberg 2007.