PETRI NET-BASED SCHEDULING OPTIMIZATION IN CNC-FMS USING GLOBAL ESTIMATION ENHANCED ANT COLONY OPTIMIZATION

Tóm tắt

This article investigates the scheduling optimization problem in a CNC-based flexible manufacturing system (FMS). A Petri net model is developed to accurately represent the dynamic, concurrent, and resource-sharing behaviors of the manufacturing system. Based on the reachability graph derived from the Petri net model, the scheduling problem is formulated as an optimal path search problem in the system state space. To efficiently explore this search space, an improved ant colony optimization (IACO) algorithm is proposed. The algorithm integrates a global estimation function, adaptive state transition rules, and a dual pheromone update mechanism to improve convergence performance and solution stability. Experimental results demonstrate that the proposed IACO algorithm significantly outperforms the classical ant colony optimization (ACO) algorithm. Specifically, the IACO reduces the average number of iterations required for convergence from 85.2 to 12.4 and achieves the optimal solution in all 20 independent runs, compared to only 15 runs for the classical ACO. The average makespan is reduced from 1256.8 to 1203.2 time units. Furthermore, in larger-scale production scenarios, the IACO achieves significant makespan reductions compared to the actual factory baseline, while also significantly reducing the convergence time to reach the optimal schedule relative to the number of iterations required by the classical ACO. These results confirm that the proposed IACO framework provides a novel and effective solution for large-scale FMS scheduling by uniquely combining Petri net reachability analysis with a global estimation-enhanced ant colony optimization, thereby addressing the limitations of local heuristic-based methods.