报告人：吴东峻

报告题目：A Control Theorist's Approach to Dynamic Optimal Transport

报告时间：2024年10月18上午9:00-9:30

报告地点：电子信息工程学院210室

报告摘要：

Over the past two decades, optimal transport has emerged as a powerful tool in artificial intelligence, driving increased interest in its application to the control community. This theory offers new perspectives for addressing control problems in areas such as large-scale multi-agent systems and filtering. In this talk, I will present my recent research on applying optimal control theory to dynamic optimal transport problems, both in continuous and discrete time settings. This includes advances in both theoretical analysis and numerical methods. I will also demonstrate that the discrete time framework is particularly well-suited for high-dimensional systems, leveraging techniques from approximate dynamic programming and reinforcement learning. At the end of this talk, some open challenges will be discussed.

在过去的二十年里，最优传输成为了人工智能等学科中一个广泛使用的理论和计算工具。近来，这一工具也引发了控制学界的持续关注，尤其是其在大规模多智能体系统和滤波等问题上潜在的理论和应用价值。本次报告，我将介绍我在动态传输方面的工作，主要是如何使用传统最优控制理论来解决动态传输问题，包括理论框架和算法。我们将会注意到，离散时间框架对于高维系统似乎更为优越，因为这个框架可以有效利用近似动态规划和强化学习等技术来近似求解最优传输。在本次报告的结尾，我将讨论一些公开问题。

报告人简介:

Dongjun Wu is currently a postdoctoral researcher at Lund University, Sweden. He obtained a double-PhD degree from University of Paris-Saclay and Harbin Institute of Technology in 2022. In 2016, he received his bachelor's degree in automation from Harbin Institute of Technology. His research focuses on the intersection of optimization and control, with a particular interest in the control and optimization of large-scale systems.

吴东峻，1993年生于重庆，2016年于哈尔滨工业大学（黑龙江）获自动化学士学位，2022年3月获哈尔滨工业大学（控制科学与工程）和巴黎萨克雷大学（自动控制）双博士学位。2022年4月起任瑞典隆德大学博士后研究员。研究兴趣包括非线性控制、大尺度系统、最优控制和最优传输。