Advanced Computational Stochastic Dynamic Programming for Continuous Time Problems

本课题致力于研究受高斯白噪声和泊松白噪声干扰的连续时间非线性动力系统的最优反馈控制的快速有效算法。泊松噪声的计算处理是这一提议的一个特别独特的特点。正在开发用于向量多处理器计算的快速算法，该算法利用问题的向量公式和算法的其他可并行化部分。另一个重要的特点是，这项工作的高级计算方面有助于消除贝尔曼的“维度诅咒”。

This project is concerned with the development of fast and efficient algorithms for the optimal feedback control of continuous time nonlinear dynamical systems, perturbed by Gaussian and Poisson white noise. The computational treatment of Poisson noise is a particularly unique feature of this proposal. Fast algorithms for vector multiprocessor computations are being developed that take advantage of vector formulation of the problem and other parallelizeable parts of the algorithm. Another important feature is that the advanced computing aspects of this work help remove Bellman's "curse of dimensionality".