Dynamic stochastic lot sizing with forecast evolution in rolling‐horizon planning

Academic approaches considering demand uncertainty in lot sizing are seldom used practice. Industry typically implements deterministic models and accounts for uncertainties by using a rolling-horizon planning framework with frequent forecast updates. This paper bridges this gap proposing stochastic lot-sizing methodology adapted to processes. Using the martingale model of evolution (MMFE), we able anticipate updates from sizing. Our formulation is extended production recourse reflect replanning flexibility planning. Extensive simulations on both synthetic real-world data show value models. Forecast reduce actual costs 14% average compared traditional The advantage depends several factors including capacity, correlation, uncertainty. Sensitivity analyses that can an additional 3% up 10% specific settings. data, provide first analysis additive multiplicative MMFE-based when true process unknown. We that, contrary existing consensus, performs more robustly than wide array problem