Mathieu Claeys

Measures and LMIs for Impulsive Nonlinear Optimal Control

This note shows how to use semi-definite programming to find lower bounds on a large class of nonlinear optimal control problems with polynomial dynamics and convex semialgebraic state constraints and an affine dependence on the control. This is done by relaxing an optimal control problem into a linear programming problem on measures, also known as a generalized moment problem. The handling of measures by their moments reduces the problem to a convergent series of standard linear matrix inequality relaxations. When the optimal control consists of a finite number of impulses, we can recover simultaneously the actual impulse times and amplitudes by simple linear algebra. Finally, our approach can be readily implemented with standard software, as illustrated by a numerical example.

Measures and LMI for impulsive optimal control with applications to space rendezvous problems

This paper shows how to find lower bounds on, and sometimes solve globally, a large class of nonlinear optimal control problems with impulsive controls using semi-definite programming (SDP). This is done by relaxing an optimal control problem into a measure differential problem. The manipulation of the measures by their moments reduces the problem to a convergent series of standard linear matrix inequality (LMI) relaxations, each giving a lower bound on the global infimum of the original problem. The case of the impulsive rendezvous of two orbiting spacecrafts is then treated. Global optimality of the solutions can be guaranteed numerically by a posteriori simulations, and we can recover simultaneously the optimal impulse time and amplitudes by simple linear algebra.

Optimal switching control design for polynomial systems: an LMI approach

We propose a new LMI approach to the design of optimal switching sequences for polynomial dynamical systems with state constraints. We formulate the switching design problem as an optimal control problem which is then relaxed to a linear programming (LP) problem in the space of occupation measures. This infinite-dimensional LP can be solved numerically and approximately with a hierarchy of convex finite-dimensional LMIs. In contrast with most of the existing work on LMI methods, we have a guarantee of global optimality, in the sense that we obtain an asympotically converging (i.e. with vanishing conservatism) hierarchy of lower bounds on the achievable performance. We also explain how to construct an almost optimal switching sequence.

Moment LMI approach to LTV impulsive control

In the 1960s, a moment approach to linear time varying (LTV) minimal norm impulsive optimal control was developed, as an alternative to direct approaches (based on discretization of the equations of motion and linear programming) or indirect approaches (based on Pontryagins maximum principle). This paper revisits these classical results in the light of recent advances in convex optimization, in particular the use of measures jointly with a hierarchy of linear matrix inequality (LMI) relaxations. Linearity of the dynamics allows us to integrate system trajectories and to come up with a simplified LMI hierarchy, where the only unknowns are moments of a vector of control measures of time. In particular, occupation measures of state and control variables do not appear in this formulation. This is in stark contrast with LMI relaxations arising usually in polynomial optimal control, where size grows quickly as a function of the relaxation order. Jointly with the use of Chebyshev polynomials (as a numerically more stable polynomial basis), this allows LMI relaxations of high order (up to a few hundreds) to be solved numerically.

Efficient upper and lower bounds for global mixed-integer optimal control

We present a control problem for an electrical vehicle. Its motor can be operated in two discrete modes, leading either to acceleration and energy consumption, or to a recharging of the battery. Mathematically, this leads to a mixed-integer optimal control problem (MIOCP) with a discrete feasible set for the controls taking into account the electrical and mechanical dynamic equations. The combination of nonlinear dynamics and discrete decisions poses a challenge to established optimization and control methods, especially if global optimality is an issue. Probably for the first time, we present a complete analysis of the optimal solution of such a MIOCP: solution of the integer-relaxed problem both with a direct and an indirect approach, determination of integer controls by means of the sum up rounding strategy, and calculation of global lower bounds by means of the method of moments. As we decrease the control discretization grid and increase the relaxation order, the obtained series of upper and lower bounds converge for the electrical car problem, proving the asymptotic global optimality of the calculated chattering behavior. We stress that these bounds hold for the optimal control problem in function space, and not on an a priori given (typically coarse) control discretization grid, as in other approaches from the literature. This approach is generic and is an alternative to global optimal control based on probabilistic or branch-and-bound based techniques. The main advantage is a drastic reduction of computational time. The disadvantage is that only local solutions and certified lower bounds are provided with no possibility to reduce these gaps. For the instances of the electrical car problem, though, these gaps are very small. The main contribution of the paper is a survey and new combination of state-of-the-art methods for global mixed-integer optimal control and the in-depth analysis of an important, prototypical control problem. Despite the comparatively low dimension of the problem, the optimal solution structure of the relaxed problem exhibits a series of bang-bang, path-constrained, and sensitivity-seeking arcs.

Modal occupation measures and LMI relaxations for nonlinear switched systems control

This paper presents a linear programming approach for the optimal control of nonlinear switched systems where the control is the switching sequence. This is done by introducing modal occupation measures, which allow to relax the problem as a primal linear programming (LP) problem. Its dual linear program of Hamilton-Jacobi-Bellman inequalities is also characterized. The LPs are then solved numerically with a converging hierarchy of primal-dual moment-sum-of-squares (SOS) linear matrix inequalities (LMI). Because of the special structure of switched systems, we obtain a much more efficient method than could be achieved by applying standard moment/SOS LMI hierarchies for general optimal control problems.

Comparison of numerical methods in the contrast imaging problem in NMR

In this article, the contrast imaging problem in nuclear magnetic resonance is modeled as a Mayer problem in optimal control. A first synthesis of locally optimal solutions is given in the single-input case using geometric methods and the HamPath software, both based on Pontryagins maximum principle. We then compare these results using direct methods implemented with the Bocop toolbox and a moment-based approach, making a first step towards global optimality.

Reconstructing trajectories from the moments of occupation measures

Moment optimization techniques have been recently proposed to solve globally various classes of optimal control problems. Since those methods return truncated moment sequences of occupation measures, this paper explores a numerical method for reconstructing optimal trajectories and controls from this data. By approximating occupation measures by atomic measures on a given grid, the problem reduces to a finite-dimensional linear program. In contrast with earlier numerical methods, this linear program is guaranteed to be feasible, no tolerance needs to be specified, and its size can be properly controlled. When combined with local optimal control solvers, this yields a powerful and flexible numerical approach for tackling difficult control problems, as demonstrated by examples.

Polynomial optimization for water networks: Global solutions for the valve setting problem

This paper explores polynomial optimization techniques for two formulations of the energy conservation constraint for the valve setting problem in water networks. The sparse hierarchy of semidefinite programing relaxations is used to derive globally optimal bounds for an existing cubic and a new quadratic problem formulation. Both formulations use an approximation for friction loss that has an accuracy consistent with the experimental error of the classical equations. Solutions using the proposed approach are reported on four water networks ranging in size from 4 to 2000 nodes and are compared against a local solver, Ipopt and a global solver, Couenne. Computational results found global solutions using both formulations with the quadratic formulation having better time efficiency due to the reduced degree of the polynomial optimization problem and the sparsity of the constraint matrix. The approaches presented in this paper may also allow global solutions to other water network steady-state optimization problems formulated with continuous variables.