Multi-objective evolutionary scheduling based on collaborative virtual workflow model and adaptive rules for flexible production process with operation reworking

Abstract

Dynamic event disturbances during production such as reworking of defective operations more challenge scheduling optimization. Efficient response to operation reworking is crucial for collaborative optimization of production quality and scheduling performance. Given the complex constraints within production processes of jobs involving key operation reworking and uncertainty of dynamic disturbances, a two-level virtual workflow modeling method and a multi-objective evolutionary scheduling algorithm with adaptive rules (MOESA-AR) based on collaborative virtual workflow framework are proposed. Virtual workflow modeling fulfills linear abstraction of tasks to be scheduled and priority constraints. Through decision arrangement, task state tracking and information feedback, collaborative virtual workflow model constitutes framework for dynamic update of task workflows and real-time response to rescheduling triggered by event disturbances. MOESA-AR adopts two-dimensional chromosome evolution and heuristic rule training to adapt to the complexity of scheduling optimization. Adaptive strategy of dispatching rules enhances the dynamic adaptability of decision-making to variable scheduling scenarios, thereby improving the dominance level of multi-objective trade-off optimization under changing demand conditions. Experiments confirm the superiority of MOESA-AR in improving scheduling performance and the effectiveness of collaborative virtual workflow scheduling framework in responding to dynamic disturbances.