An event-driven hybrid rescheduling approach for integrated process planning and scheduling considering stochastic rework.

While the static integrated process planning and scheduling (IPPS) problem is theoretically well-established, its practical application is limited in unpredictable manufacturing environments demanding dynamic adaptability. This paper proposes a dynamic IPPS problem considering stochastic rework (IPPS-SR), whose solution optimizes product quality and scheduling performance when imperfect items require reprocessing. We first formulate a mathematical optimization model for IPPS-SR that minimizes makespan and schedule instability, and then present an event-driven hybrid rescheduling approach featuring two key innovations: (1) a hybrid strategy that integrates right-shift scheduling with a multi-objective reinforcement learning-guided adaptive large neighborhood search (MORL-ALNS) algorithm, achieving an effective trade-off between computational efficiency and solution quality; and (2) a set of problem-specific operators, including five destroy and four repair operators, that enhance the search efficacy of the MORL-ALNS framework. Experimental results on 24 adapted benchmark instances indicate that the proposed hybrid approach effectively generates high-quality rescheduling schemes for the IPPS-SR problem. Specifically, the RL-guided mechanism increases the number of non-dominated solutions by over 80% on average compared to the baseline ALNS. Comprehensive experiments with other widely used multi-objective algorithms further demonstrate that MORL-ALNS achieves superior Hypervolume (HV) values in 22 out of 24 instances and lower Inverted Generational Distance (IGD) values in 23 out of 24 instances.