Genetic optimization of energy-and failure-aware continuous production scheduling in pasta manufacturing

Abstract

Energy and failure are separately managed in scheduling problems despite the commonalities between these optimization problems. In this paper, an energy-and failure-aware continuous production scheduling problem (EFACPS) at the unit process level is investigated, starting from the construction of a centralized combinatorial optimization model combining energy saving and failure reduction. Traditional deterministic scheduling methods are difficult to rapidly acquire an optimal or near-optimal schedule in the face of frequent machine failures. An improved genetic algorithm (IGA) using a customized microbial genetic evolution strategy is proposed to solve the EFACPS problem. The IGA is integrated with three features: Memory search, problem-based randomization, and result evaluation. Based on real production cases from Soubry N.V., a large pasta manufacturer in Belgium, Monte Carlo simulations (MCS) are carried out to compare the performance of IGA with a conventional genetic algorithm (CGA) and a baseline random choice algorithm (RCA). Simulation results demonstrate a good performance of IGA and the feasibility to apply it to EFACPS problems. Large-scale experiments are further conducted to validate the effectiveness of IGA.