A multi-agent system simulation of job shop scheduling with human consideration: A comparative analysis of AGVs and AIVs

Abstract

The manufacturing landscape is undergoing a paradigm shift towards Industry 5.0, emphasizing human-centricity in a collaborative environment between humans and robots. In this context, the job shop scheduling problem (JSSP) remains a critical aspect of workshop management, optimizing task sequencing to minimize the overall completion time (makespan). Traditionally, autonomous and guided vehicles (AGVs) have been employed for tasks transportation within workshops. However, their limited flexibility and reliance on dedicated pathways hinder seamless integration with human workers. Autonomous and intelligent vehicles (AIVs) emerge as a promising alternative, offering enhanced adaptability and the ability to navigate alongside humans in shared spaces. Despite the emphasis on human-centricity as a foundation of manufacturing in Industry 5.0, research explicitly considering human factors in JSSPs remains scarce. Hence, this study delves into the transition from AGVs to AIVs within the context of JSSP, evaluating their impact on the makespan and the influence of human presence on AIV operations. Employing a multi-agent system (MAS)-based simulation approach, we model the stochastic human behaviour using discrete-time Markov chains (DTMC) and investigate the impact of the interactions between AIVs and human workers within shared movement corridors. Our findings highlight the potential of AIVs to enhance workshop efficiency over AGVs, and their effectiveness in human-centred manufacturing policy, paving the way for more collaborative and adaptable production systems.