Deadlock-free scheduling method for automated manufacturing systems with limited central buffers

An efficient algorithm based on graph-theoretic approach and genetic algorithm for finding optimal deadlock-free schedules in the automated manufacturing systems with limited central buffers is presented. The scheduling problem introduced In this work involves sequencing the operations on resources to avoid the deadlock situation and to minimize the makespan. With the help of digraph models of automated manufacturing systems, deadlock avoidance policy is proposed to supervise the release and acquisition of resources. In order to satisfy the deadlock free constraint of chromosomes, the deadlock avoidance policy is embedded into the decoding procedure of genetic algorithm and after adaptation of gene sequences, the feasibility of solutions can always be guaranteed for the peculiarity of the scheduling problem. With this deadlock free scheduling method, the optimal deadlock free schedules can be got under the restriction of different buffer capacities. The efficiency of deadlock free scheduling algorithm is proved by simulation results.