Multi-load AGVS deadlock prevention task scheduling method based on improved imperialist competition algorithm

Abstract

Vehicle production involves a complex process and fast production pace, necessitating timely material delivery. The task scheduling problem of multi-load Automated Guided Vehicle System (AGVS) for auxiliary material distribution in automotive production workshops presents multiple optimization objectives and complex constraints. To enhance on-time delivery of auxiliary materials, this paper proposes a multi-load AGVS deadlock prevention task scheduling method based on improved Imperialist Competition Algorithm (ICA). First, a mathematical model for multi-load AGVS task scheduling is established, aiming to minimize task delivery distance and maximize the remaining time before the production line shuts down due to material shortages. By analyzing the conditions triggering deadlocks in multi-load AGVS, a deadlock prevention constraint based on the remaining trailer capacity of the buffer area is incorporated into the mathematical model. Second, an improved ICA (IICA) based deadlock prevention task scheduling method is designed. To enhance the initial national population quality of the IICA, a heuristic scheduling rule library is constructed to generate high-quality countries. An improved differential evolution algorithm is introduced during the assimilation process to improve convergence speed. Finally, a simulation platform for multi-load AGVS auxiliary material distribution is established. Experimental results indicate that the designed deadlock avoidance strategy effectively prevents deadlock occurrences while enhancing system productivity across all six algorithms. Compared to the other five algorithms, the proposed IICA achieves the highest unit hour production capacity, on-time task completion rate, and production line start-up rate, while maintaining the lowest average task execution time.