Stochastic control of uncertain systems: a model in which the control increases uncertainty

• Grant number: 16/02208-6
• Support Opportunities: Scholarships in Brazil - Master
• Start date: April 01, 2016
• End date: June 30, 2018
• Field of knowledge: Engineering - Electrical Engineering
• Principal Investigator: João Bosco Ribeiro do Val
• Grantee: Vinicius Lima Silva
• Host Institution: Faculdade de Engenharia Elétrica e de Computação (FEEC). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Associated scholarship(s): 17/10340-4 - Control of uncertain stochastic systems: the CVIU approach and the H infinity control, BE.EP.MS

Abstract

The research project is based on the analysis of stochastic systems for which control variations increase state uncertainty. This class was named CVIU systems and, as shown in previous papers, a technique allowing the design of controllers which, in face of high uncertainty regarding the system's model, offers a better performance than that of classical controllers, was developed. The basic idea underlying the controller design consists of the addition of an extra random term modulated by the control action intensity to the system model. Even though promising results have been found so far, calculating the optimal controller is still a numerically challenging task. Especially difficult is the resolution of a highly nonlinear differential equation derived from the Hamilton-Jacobi-Bellman equation obtained via the dynamic programming method. The solution of this equation, however, is essential to the calculation of the optimal control policy. Therefore, the study of numerical methods which may offer a more efficient form of solving this nonlinear differential equation is one of the main foci of this project. Also, an extension of the CVIU model is proposed in this work. Another aim of the project is to analyze whether a more general description of the system modeling uncertainty offers an improvement in the controller performance. In addition, the application of the CVIU model to the analysis and control of systems with poorly known dynamics described by growth equations poorly known parameters will be studied. Typical scenarios of aplications are tumor growth control, fish harvesting and macroeconomic policies.

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

SILVA, VINICIUS L.; DO VAL, JOAO B. R.; SOUTO, RAFAEL F.; IEEE. A Stochastic Approach for Robustness: a H-2-norm comparison. 2017 IEEE 56TH ANNUAL CONFERENCE ON DECISION AND CONTROL (CDC), v. N/A, p. 6-pg., 2017-01-01. (16/02208-6, 16/08645-9)

Academic Publications

(References retrieved automatically from State of São Paulo Research Institutions)

SILVA, Vinicius Lima. Normas e estabilidade para modelos estocásticos cuja variação do controle e do estado aumentam a incerteza. 2018. Master's Dissertation - Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Elétrica e de Computação Campinas, SP.