Retour au site du LAAS-CNRS

Abstract

The numerical range of a matrix is studied geometrically via the cone of positive semidefinite matrices (or semidefinite cone for short). In particular it is shown that the feasible set of a two-dimensional linear matrix inequality (LMI), an affine section of the semidefinite cone, is always dual to the numerical range of a matrix, which is therefore an affine projection of the semidefinite cone. Both primal and dual sets can also be viewed as convex hulls of explicit algebraic plane curve components. Several numerical examples illustrate this interplay between algebra, geometry and semidefinite programming duality. Finally, these techniques are used to revisit a theorem in statistics on the independence of quadratic forms in a normally distributed vector.

Mots-Clés / Keywords

Abstract

This paper presents an inverse optimality method to solve the Hamilton-Jacobi-Bellman equation for a class of nonlinear problems for which the cost is quadratic and the dynamics are affine in the input. The method is inverse optimal because the running cost that renders the control input optimal is also explicitly determined. One special feature of this work, as compared to other methods in the literature, is the fact that the solution is obtained directly for the control input. The value function can also be obtained after one solves for the control input. Furthermore, a Lyapunov function that proves at least local stability of the controller is also obtained. In this regard the main contribution of this paper can be interpreted in two different ways: offering an analytical expression for Lyapunov functions for a class of nonlinear systems and obtaining an optimal controller for the same class of systems using a specific optimization functional. We also believe that an additional contribution of this paper is to identify explicit classes of systems and optimization functionals for which optimal control problems can be solved analytically. In particular, for second order systems three cases are identified: i) control input only as a function of the second state variable, ii) control input affine in the second state variable when the dynamics are affine in that variable and iii) control input affine in the first state variable when the dyamics are affine in that variable. The relevance of the proposed methodology is illustrated in several examples, including the Van der Pol oscillator, mass-spring systems and vehicle path following.

Abstract

Given a polynomial $x \in {\mathbb R}^n \mapsto p(x)$ in $n=2$ variables, a symbolic-numerical algorithm is first described for detecting whether the connected component of the plane sublevel set ${\mathcal P} = \{x : p(x) \geq 0\}$ containing the origin is rigidly convex, or equivalently, whether it has a linear matrix inequality (LMI) representation, or equivalently, if polynomial $p(x)$ is hyperbolic with respect to the origin. The problem boils down to checking whether a univariate polynomial matrix is positive semidefinite, an optimization problem that can be solved with eigenvalue decomposition. When the variety ${\mathcal C} = \{x : p(x) = 0\}$ is an algebraic curve of genus zero, a second algorithm based on B\'ezoutians is proposed to detect whether $\mathcal P$ has an LMI representation and to build such a representation from a rational parametrization of $\mathcal C$. Finally, some extensions to positive genus curves and to the case $n>2$ are mentioned.

Mots-Clés / Keywords

Abstract

This paper presents a projection-based approach for solving conic feasibility problems. To
nd a point in the intersection of a cone and an ane subspace, we simply project a point onto this intersection. This projection is computed by dual algorithms operating a sequence of projections onto the cone, and generalizing the alternating pro jection method. We release an easy-to-use Matlab package implementing an elementary dual projection algorithm. Numerical experiments show that, for solving some semide
nite feasibility problems, the package is competitive with sophisticated conic programming software. We also provide a particular treatment of semide
nite feasibility problems modeling polynomial sum-of-squares decomposition problems.

Abstract

Given a basic compact semi-algebraic set $K$ in $R^n$ we introduce a methodology that provides a sequence converging to the volume of $K$. This sequence is obtained from optimal values of a hierarchy of either semidefinite or linear programs. Not only the volume but also every finite vector of moments of the probability measure uniformly distributed on $K$ can be approximated as closely as desired, and so permits to approximate the integral on $K$ of any given polynomial; extension to integration against some weight functions is also provided. Finally, some numerical aspects are discussed.

Mots-Clés / Keywords

Abstract

We describe a major update of our Matlab freeware GloptiPoly for parsing generalized problems of moments and solving them numerically with semidefinite programming.

Abstract

POCP is a new Matlab package running jointly with GloptiPoly 3 and, optionally, YALMIP. It is aimed at nonlinear optimal control problems for which all the problem data are polynomial, and provides an approximation of the optimal value as well as some control policy. Thanks to a user-friendly interface, POCP reformulates such control problems as generalized problems of moments, in turn converted by GloptiPoly 3 into a hierarchy of semidefinite programming problems whose associated sequence of optimal values converges to the optimal value of the polynomial optimal control problem. In this paper we describe the basic features of POCP and illustrate them with some numerical examples.

Mots-Clés / Keywords