Decomposition of timed automata for solving jobshop scheduling problems

We propose a decomposition algorithm for solving jobshop scheduling problems represented by timed automata (TA). Jobshop scheduling problem is formulated as an optimal state transition problem for the parallel composition of submodels such as jobs and machines. The entire model is decomposed into tractable submodels. The decomposable conditions are provided. In the proposed algorithm, a feasible solution for the entire model is derived through the repeated computation of solving subproblem for each decomposed element. Computational experiments demonstrate the effectiveness of the proposed algorithm for jobshop instances.