A unified robust stochastic programming approach for integrated production and distribution scheduling problem for seru production system under uncertainty

This study addresses the seru production system production and distribution scheduling (SPSPDS) problem under uncertainty by integrating production and distribution scheduling decisions within a lot streaming context. A unified robust stochastic programming (URSP) model is proposed to minimize the maximum lead time (MLT) by simultaneously accounting for two types of demand uncertainty: known–unknown uncertainty, handled through a stochastic programming component, and unknown–unknown uncertainty, addressed via a robust optimization component to enhance risk aversion. This flexible modeling structure enables the representation of three real-world operational modes for seru systems: regular, seasonal, and emergent. The proposed approach is well aligned with the principles of Just-In-Time Organization Systems (JIT-OS), ensuring that serus are available at the right time, place, and capacity, even under deep uncertainty. Multiple solution approaches are developed based on an artificial bee colony algorithm (ABCA), adapted to different combinations of sublot sequencing and vehicle assignment strategies. A full factorial design of experiments is employed to systematically examine the effects of controllable factors, demand distribution characteristics, and robustness levels. Computational results demonstrate that selecting appropriate sublot sequencing and vehicle assignment strategies according to demand characteristics significantly improves operational efficiency, reduces lead times, and enhances customer satisfaction. Overall, this study provides theoretical, methodological, and managerial contributions by introducing a novel URSP-based approach for solving the SPSPDS problem.