NSF Workshop: Bridging the Gap Between Operations Research and Approximate Dynamic Programming, April 16-20, 2005, in Cancun, Mexico

长期以来，动态规划一直有望解决一些更具挑战性的决策问题。 然而，这些方法都遭受了众所周知的“维数灾难”，这基本上限制了它们的适用性，决策问题的几个状态变量。 在近似这些决策问题的最新进展导致了被称为近似动态规划的方法。 这些方法与其他一些方法密切相关，如线性规划和随机规划。 本次研讨会将探讨这些方法之间的联系，以及在运输，电力系统等方面出现的几个重要应用。 该研讨会将汇集来自人工智能，神经网络和运筹学等多个学科的研究人员和学生，专注于近似动态规划及相关领域的新突破。

Dynamic programming has long held the promise of solving some of the more challenging decision problems. However, these approaches have suffered from the well known "curse of dimensionality" which essentially limits their applicability to decision problems with a few state variables. Recent advances in approximating these decision problems have led to methods known as approximate dynamic programming. These methods are intimately related to several other methods such as linear programming, and stochastic programming. This workshop will explore connections among these methods, as well as several important applications arising in transportation, power systems, and others. This workshop will bring together researchers and students from several disciplines, such as artificial intelligence, neural networks, and operations research to focus on new breakthroughs in approximate dynamic programming, and related areas.