Scheduling Optimization Based on Flexible Manufacturing Systems Modeled With Timed Petri Nets

This paper deals with the problem of scheduling optimization in flexible manufacturing systems modeled with timed Petri nets by using the ant colony algorithm and combining the backtrack algorithm. In the flexible manufacturing process, potential resource failures may lead to deadlock issues in scheduling, which may result in inefficient resource allocation and even affect the completion of production tasks. To this end, the backtrack ant colony algorithm is proposed to optimize scheduling in the timed Petri net model and to overcome the slow convergence and susceptibility to local optimal inherent in existing algorithms. Specifically, in the initial state, the regular ant colony algorithm is used for path finding, but when ants encounter dead situations, i.e., they cannot proceed, the backtrack algorithm revives them and allows to choose new paths to ensure the discovery of optimal scheduling results. Finally, the algorithm is tested by using a timed Petri net that may enter deadlocks. Compared to conventional algorithms, the backtrack ant colony algorithm demonstrates shorter run times and is capable of providing high-quality scheduling solutions and accelerating convergence on complex network models.