报告题目: A Distributionally Robust Approach to Chance Constrained Stochastic Model Predictive Control Problems with Unbounded Disturbance
报告时间: 15:00-17:00， 2018年11月13
Abstract: Model predictive control (MPC) has many real world applications, such as in supply chain, inventory control and process control. It has been proved to be effective by being applied in the industry. In this talk, we propose a state feedback control for stochastic model predictive control (SMPC) in the discrete linear systems with unbounded noise. Participially, chance constraints on the state and control are imposed. The proposed method is based on the deterministic reformulation of these constraints. Different from the state of the art method which usually applies the Chebyshev-Cantelli inequality to approximate the chance constraints with intractable reformulation, the proposed algorithm provides an equivalent tractable alternative by utilizing a distributionally robust framework. Despite this, there is no need to update the state covariance matrix at each time instant with our proposed scheme. The recursive feasibility, stability and convergence of the algorithm are also proved. Numerical results are provided to demonstrate the effectiveness of the proposed SMPC algorithm.
报告人简介：李彬，四川大学电气信息学院担任特聘研究员。 2011年在澳大利亚科廷大学获得数学与统计系博士学位。2012年至2017在西澳大利亚大学电气电子与计算机工程系从事和澳大利亚科廷大学数学与统计系从事研究工作。目前， 李彬研究员主要从事数值优化在控制理论与工程和信号处理等领域中应用，共发表SCI检索期刊20余篇，其中包括IEEE Transactions on Signal Processing和IEEE Transactions on Communications等多个IEEE顶级杂志。此外，李研究员担任过多个SCI杂志（如Journal of Industrial Management and Optimization）客座主编，并且在包括Mathematical Programming，Automatica和IEEE Transactions on Signal Processing在内的10多个杂志担任常年审稿人。