Multi-stage Stochastic Programming Models in Production Planning

Production planning is a key area of operations management. An important methodology for production planning is mathematical programming. Traditional mathematical programming models for production planning are deterministic, and cannot provide robust production plans in the presence of uncertainty. As such, deterministic planning models may yield unsatisfactory decisions. Stochastic programming, an active branch of mathematical programming dealing with optimization problems involving uncertain data, has seen several successful applications in production planning. Unlike alternative approaches to decision making under uncertainty, such as Markov decision processes, stochastic programming requires few assumptions on the underlying stochastic processes and allows for modeling of complicated decision structures. On the other hand, stochastic programming assumes finite number of stages and exogenous uncertainties. With recent increase in computational power and algorithmic developments, the limitations of stochastic programming arising from computational difficulties have been relieved to a large extent. Nowadays, good production planning is a considered as one of the reason for improvement in production and many studies have been conducted in order to identify the models of production planning. The main purpose of this research is to study multi-stage stochastic programming models in production planning.