Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.


Stochastic Reachability using Polynomial Optimization


D. Drzajic

Master Thesis, HS15 (10358)

In the Automatic Control Laboratory IfA at ETH it has been studying reachability objectives for stochastic dynamical systems, formulated as optimal control problems. It has been shown that the probability that a controlled stochastic dynamical system reaches a given target set while avoiding another set can be maximized via a dynamic programming recursion. It has been implemented different numerical solutions to this recursion based on space gridding and radial basis function approximation. Recent advances in polynomial optimization allow solving optimal control dynamic recursions for systems with polynomial dynamics. One of the main challenges in this approach is the representation of system dynamics in the polynomial framework and the degree of approximation introduced by doing so. Moreover, the size of problems that can be handled with existing software varies depending on the complexity of the original control problem. In this project we will investigate the possibility of the polynomial optimization framework and software tools to solve discrete time stochastic reachability problems. Our goal is to investigate the quality of the solutions obtained using polynomial optimization tools and compare to the existing approximations obtained via space gridding and radial basis function approximation. We will carry out a comparison based on accuracy, scalability and total computational time.

Supervisors: Nikolaos Kariotoglou, John Lygeros


Type of Publication:

(12)Diploma/Master Thesis

J. Lygeros

File Download:

Request a copy of this publication.
(Uses JavaScript)
% Autogenerated BibTeX entry
@PhdThesis { Xxx:2016:IFA_5375
Permanent link