Transgenic Algorithm Applied to the Job Shop Rescheduling Problem

Abstract

Context: Job sequencing has been approached from a static perspective, without considering the occurrence of unexpected events that might require modifying the schedule, thereby affecting its performance measures. Method: This paper presents the development and application of a genetic algorithm to the Job Shop Rescheduling Problem (JSRP), a reprogramming of the traditional Job Shop Scheduling Problem. This novel approach seeks to repair the schedule in such a way that theoretical models accurately represent real manufacturing environments. Results: The experiments designed to validate the algorithm aim to apply five classes of disruptions that could impact the schedule, evaluating two performance measures. This experiment was concurrently conducted with a genetic algorithm from the literature in order to facilitate the comparison of results. It was observed that the proposed approach outperforms the genetic algorithm 65% of the time, and it provides better stability measures 98% of the time. Conclusions: The proposed algorithm showed favorable outcomes when tested with well-known benchmark instances of the Job Shop Scheduling Problem, and the possibility of enhancing the tool's performance through simulation studies remains open.