Operation decomposition and statistical bottleneck machine identification for large-scale job shop scheduling

An decomposition-based optimization algorithm is presented for scheduling large-scale job shops with the objective of minimizing total weighted tardiness. In each iteration, we first define a new subproblem which contains a subset of operations selected from the original problem, and then we solve this newly defined subproblem using a genetic algorithm. Before each subproblem is solved, bottleneck machines are identified by a statistical method to reflect the characteristic information concerning the impending subproblem. Then, the characteristic information is used to determine the encoding scheme for the genetic algorithm. Numerical computational results show that the proposed algorithm is effective for solving large-scale scheduling problems.