A hybrid approach to large-scale job shop scheduling

A decomposition based hybrid optimization algorithm is presented for large-scale job shop scheduling problems in which the total weighted tardiness must be minimized. In each iteration, a new subproblem is first defined by a simulated annealing approach and then solved using a genetic algorithm. We construct a fuzzy inference system to calculate the jobs' bottleneck characteristic values which depict the characteristic information in different optimization stages. This information is then utilized to guide the process of subproblem-solving in an immune mechanism in order to promote the optimization efficiency. Numerical computational results show that the proposed algorithm is effective for solving large-scale scheduling problems.