Advances in stochastic dynamic programming for operations management

En savoir plus sur le livre

Many tasks in operations management involve complex optimization problems that require sequential decision-making over time, which can be effectively modeled using dynamic programming. However, real-world scenarios often present complexities that conventional techniques struggle to address, including large state spaces, extensive action sets, non-convex objective functions, and uncertainty. This work explores three intricate problems in operations management, employing newly developed stochastic dynamic programming techniques. The first problem focuses on optimally scheduling demand units in an energy transmission network to minimize the total cost of electric energy supply amid demand and generation uncertainties. The second problem addresses the integrated investment and operational strategies for a network of battery swap stations, considering uncertain demand and fluctuating energy prices. The third problem involves inventory control for a multi-channel retailer, where optimal replenishment policies with simple structures are established. The book introduces efficient approximation techniques based on approximate dynamic programming (ADP) and enhances existing proximal point algorithms for stochastic scenarios, making these methods applicable to a diverse range of high-dimensional dynamic programming challenges.