Improved Prairie Dog Optimization Algorithm for Online Detection of Automated Scheduling Deadlock-Prone Manufacturing System

Production scheduling involves the determination of resources as well as the sequence and timing with which jobs are done to accomplish one or more objectives and it a critical aspect of the industrial manufacturing process. Due to limited space available in the production floor or for certain technological reasons, blocking restrictions are mostly implemented in the various production floors that are normally found in industries for instance, electronic manufacturing shops, robotic cells or iron and steel industries. Indeed, the research of this work is proposed on the scheduling problems of a type of Petri nets, namely the so-called S3PRs in the context of for modelling deadlock-prone Automated Manufacturing Systems (AMSs) in online detection. It also proposes an Improved Prairie Dog Optimization (IPDO) Algorithm to reduce the makespan value. First, by extending the concepts of place-timed Petri networks to AMSs, the timed state space (TSS) including the timed state is defined. Thus, to enhance the solving performance of various scale AMSs’ scheduling problems, TSS is compressed, and, based on this, IPDO is developed for the CSS’s search. In the IPDO model, the transition policy of duplicate detection is introduced to guarantee that only one path is reserved for each state in CSS. A pruning policy is presented to prune out those kinds of states which cannot win the game in order to search a small part of CSS, and three different cost estimation functions are discussed for states. Due to the properties of the search algorithm used, a new search strategy, which is a combination of the above techniques, is adopted in order to increase the efficiency and produce better solutions. The proposed method attained the 465 s in 10 instances when comparatively less time with other existing approaches.