永利集团3044官网欢迎您理学院“明理志远”学术讲坛(二十七)

发布时间:2023-06-13浏览次数:679文章来源:理学院

  题 目: Perimeter Control With State-Dependent Delays: Optimal Control Model and Computational Method

  时 间:2023年6月15日12:00-14:00

  地 点:理学院405

  主讲人:袁金龙 教授

  袁金龙,博士,大连海事大学副教授,硕士生导师,主要从事不确定性的分布鲁棒优化与最优控制理论、算法及应用方面的研究。近几年来在国内外重要优化与控制期刊上发表或录用学术论文39篇,其中SCI检索36篇。主持国家自然科学青年基金等科研项目4项。目前担任期刊 Big Data and Cloud Innovation 的编委、美国数学会《数学评论》评论员,曾被邀请为 Automatica等12个国际 SCI 期刊审稿。入选大连海事大学“星海工程”教师培养计划第3层次和第四层次各1次;获得“辽宁省百千万人才工程”万人层次;获得辽宁省自然科学学术成果奖和大连市优秀科技成果奖。

  报告简介:

  Perimeter control is to manipulate traffic flows in different regions through adjusting traffic signals at the border of the regions. Traditionally, the complete trips in a region are assumed to be dependent on the current accumulated vehicles in this region. This assumption is invalid because the vehicle needs time to complete its trip in a region. In order to solve this shortcoming, some recent studies have introduced a timedelay dynamical system to describe the dynamical behaviour of the accumulated vehicles in a region. However, in these studies, the perimeter control problem is formulated as a tracking problem with a given optimal reference point. In reality, such an optimal reference point is unavailable in advance. This paper will fill this gap through formulating perimeter control as an optimal control problem governed by a state-dependent time delay system. The control parametrization technique and an exact penalty method are introduced to solve such a challenging optimal control problem. Model predictive control is applied to obtain close-loop solution through solving a series of online optimal control problems. Some experiments are performed to demonstrate the effectiveness of our method.


主办单位:永利集团3044官网欢迎您理学院


最新更新