A knowledge-driven memetic algorithm for energy-aware flexible job shop scheduling with limited AGV transportation

Abstract Flexible Job Shop Scheduling Problems (FJSP) traditionally assumeinfinite transportation resources or simplified transportation constraints.With the rise of intelligent manufacturing,Automated Guided Vehicles (AGVs) have emerged as essential transport resources due to their high flexibility and autonomy.The limited availability of AGVs can significantly impact overall production efficiency.Meanwhile, growing concerns about energy consumption and environmental sustainabilityunderscore the necessity of incorporating energy-related objectives into scheduling decisions.In this context, this paper addresses the Energy-aware FJSP with limited AGVs (EFJSP-AGV).A multi-objective mixed-integer programming (MMIP) model is developedto simultaneously minimize makespan and total energy consumption (TEC).To efficiently tackle this challenging problem,a knowledge-driven memetic algorithm (KDMA) is proposed.Specifically, an integrated initialization approach is devisedto efficiently generate promising initial solutions.Furthermore, a knowledge-driven variable neighborhood search (VNS) tailored tothe characteristics of the problem is developedto enhance the exploitation within the solution space.Additionally, effective energy-aware strategies for reducing energy consumption are incorporatedto achieve lower total energy usage.Experimental results indicate that the proposed KDMA outperforms comparison algorithms,validating its effectiveness in solving the EFJSP-AGV.