Short-Term Scheduling of Multi-Product Batch Plants with Sequence-Dependent Changeovers Using Timed Automata Models

摘要

In this contribution we propose an approach to model and solve scheduling problemsrnwith sequence-dependent changeover procedures using timed automata (TA). Thernprocessing units and the recipes are modeled as sets of interacting TA in a modularrnfashion. The setup and changeover procedures are modeled explicitly as operations inrnthe recipe. The problem modeled as TA is then solved to derive schedules byrnperforming a cost-optimal reachability analysis starting from an initial location wherernno recipe operation is started to a target location where all required demands are met.rnThe proposed approach is applied to two case studies and the results are presented. Therncomparison of the experimental results from the TA-based approach with establishedrnMILP techniques shows that the proposed approach is efficient in the sense that optimalrnschedules are found within a limited computation time that are comparable to or betterrnthan those obtained with the MILP formulations and the solvers considered.