Exact slow–fast decomposition of the nonlinear singularly perturbed optimal control problem

We study the infinite horizon nonlinear quadratic optimal control problem for a singularly perturbed system, which is nonlinear in both, the slow and the fast variables. It is known that the optimal controller for such problem can be designed by finding a special invariant manifold of the corresponding Hamiltonian system. We obtain exact slow–fast decomposition of the Hamiltonian system and of the special invariant manifold into the slow and the fast ones. On the basis of this decomposition we construct high-order asymptotic approximations of the optimal state-feedback and optimal trajectory.     

Averaging of three time scale singularly perturbed control systems

We investigate the asymptotic properties of singularly perturbed control systems with three time scales. We apply the averaging method to construct a limiting system for the slowest motion in the form of a differential inclusion. Sufficient conditions for the uniform convergence of the slowest trajectories are given.     

The explicit linear quadratic regulator for constrained systems

For discrete-time linear time invariant systems with constraints on inputs and states, we develop an algorithm to determine explicitly, the state feedback control law which minimizes a quadratic performance criterion. We show that the control law is piece-wise linear and continuous for both the finite horizon problem (model predictive control) and the usual infinite time measure (constrained linear quadratic regulation). Thus, the on-line control computation reduces to the simple evaluation of an explicitly defined piecewise linear function. By computing the inherent underlying controller structure, we also solve the equivalent of the Hamilton-Jacobi-Bellman equation for discrete-time linear constrained systems. Control based on on-line optimization has long been recognized as a superior alternative for constrained systems. The technique proposed in this paper is attractive for a wide range of practical problems where the computational complexity of on-line optimization is prohibitive. It also provides an insight into the structure underlying optimization-based controllers.     

A convex optimization approach for solving the single-vehicle cyclic inventory routing problem

This paper investigates the mathematical structure of the Single-Vehicle Cyclic Inventory Routing Problem (SV-CIRP). The SV-CIRP is an optimization problem consisting of finding a recurring distribution plan, from a single depot to a selected subset of retailers, that maximizes the collected rewards from the visited retailers while minimizing transportation and inventory costs. It appears as fundamental building block for all variants of the cyclic inventory routing problem (CIRP). One of the main complications in developing solution methods for the SV-CIRP using the current formulations is the non-convexity of the objective function. We demonstrate how the problem can be reformulated so that its continuous relaxation is a convex optimization problem. We further examine its mathematical properties and compare our findings with statements previously done in literature. Based of these findings we propose an algorithm that solves the SV-CIRP more effectively. We present experimental results on well-known benchmark instances, for which we are able to find optimal solutions for 22 out of 50 instances and obtained new best known solutions to 23 other instances.     

output feedback control design for uncertain fuzzy singularly perturbed systems: an LMI approach

This paper examines the problem of designing a robust H_∞ output feedback controller for a class of singularly perturbed systems described by a Takagi–Sugeno fuzzy model. Based on a linear matrix inequality (LMI) approach, LMI-based sufficient conditions for the uncertain singularly perturbed nonlinear systems to have an H_∞ performance are derived. To eliminate the ill-conditioning caused by the interaction of slow and fast dynamic modes, solutions to the problem are presented in terms of LMIs which are independent of the singular perturbation . The proposed approach does not involve the separation of states into slow and fast ones and it can be applied not only to standard, but also to nonstandard singularly perturbed nonlinear systems. A numerical example is provided to illustrate the design developed in this paper.     

The disturbance decoupling problem in decentralized linear multi-variable systems by local dynamic feedback

Following the works of Hu and Zheng [7] on the problem of disturbance decoupling in decentralized systems with non-dynamic feedback (DDPP), this paper studies the problem of disturbance decoupling in decentralized linear multi-variable systems with one or two local dynamic feedbacks (abbreviated as DDDPDC1 or DDDPDC2 respectively). A counter-example is given to indicate that the main results of Cury [1] are incorrect. Developing the concepts such as structure subspace pair and structure subspace vector, the necessary and sufficient conditions of DDDPDC1 and DDDPDC2 using low order compensators is derived.     

Implicit-system approach to the robust stability for a class of singularly perturbed linear systems

Asymptotic stability and the complex stability radius of a class of singularly perturbed systems of linear differential-algebraic equations (DAEs) are studied. The asymptotic behavior of the stability radius for a singularly perturbed implicit system is characterized as the parameter in the leading term tends to zero. The main results are obtained in direct and short ways which involve some basic results in linear algebra and classical analysis, only. Our results can be extended to other singular perturbation problems for DAEs of more general form.     

An observer-based decentralized tracker for sampled-data systems: an evolutionary programming approach

Decentralized control is a practical control methodology for large-scale multivariable systems. This paper presents a LQR design methodology to design a state-feedback decentralized high-gain analog controller, which gives the desired decentralized performance of the controlled analog system. Then, a prediction-based decentralized low-gain digital controller is developed from the decentralized high-gain analog controller for the hybrid controlled system. As a result, the complexity and cost of hardware implementation of the controller can be significantly reduced. In order to improve the performance of the decentralized hybrid system, the evolutionary programming (EP) is employed to tune the observer-based decentralized tracker. Some examples are presented to illustrate the developed design methodology.     

Controller design for a class of nonlinear systems with input saturation using convex optimization

In this paper, we propose a new design strategy for nonlinear systems with input saturation. The resulting nonlinear controllers are locally asymptotically stabilizing the origin. The proposed methodology is based on exact feedback linearization which is used to reformulate the nonlinear system as a linear system having state-dependent input saturation. Linear saturating state feedback controllers and soft variable-structure controllers are developed based on this system formulation. The resulting convex optimization problems can be written in terms of linear matrix inequalities and sum of squares conditions for which efficient solvers exist. Polynomial approximation based on Legendre polynomials is used to extend the methodology to a more general class of nonlinear systems. To demonstrate the benefit of this design method, a stabilizing controller for a single link manipulator with flexible joint is developed.     

Regulator problem for linear systems with constraints on control and its increment or rate

This paper discusses the problem of constraints on both control and its rate or increment, for linear systems in state space form, in both the continuous and discrete-time domains. Necessary and sufficient conditions are derived for autonomous linear systems with constrained state increment or rate (for the continuous-time case), such that the system evolves respecting incremental or rate constraints. A pole assignment technique is then used to solve the inverse problem, giving stabilizing state feedback controllers that respect non-symmetrical constraints on both control and its increment or rate. An illustrative example shows the application of the method on the double integrator problem.     

The discrete minimum principle for quadratic spline discretization of a singularly perturbed problem

We consider a singularly perturbed boundary value problem with two small parameters. The problem is numerically treated by a quadratic spline collocation method. The suitable choice of collocation points provides the discrete minimum principle. Error bounds for the numerical approximations are established. Numerical results give justification of the parameter-uniform convergence of the numerical approximations.     

The relaxed optimal control problem for Mean-Field SDEs systems and application

The present study deals with a new approach of optimal control problems where the state equation is a Mean-Field stochastic differential equation, and the set of strict (classical) controls need not be convex and the diffusion coefficient depends on the term control. Our consideration is based on only one adjoint process, and the necessary conditions as well as a sufficient condition for optimality in the form of a relaxed maximum principle are obtained, with application to Linear quadratic stochastic control problem with mean-field type.     

一类非线性时滞互联系统模糊分散输出反馈控制

对于一类状态不可测非线性互联时滞系统,给出一种基于观测器的模糊分散输出反馈控制方法.首先采用模糊T-S模型对非线性互联时滞系统进行模糊建模,在此基础上给出了模糊分散观测器和基于观测器的模糊分散输出控制器的设计.应用李亚普诺夫函数法和线性矩阵不等式方法给出了模糊分散控制系统稳定的充分条件.仿真结果进一步验证了所提出的模糊分散控制方法的有效性.     

Nonlinear robust H-infinity filtering for a class of uncertain systems via convex optimization

A new approach for robust H-infinity filtering for a class of Lipschitz nonlinear systems with time-varying uncertainties both in the linear and nonlinear parts of the system is proposed in an LMI framework. The admissible Lipschitz constant of the system and the disturbance attenuation level are maximized simultaneously through convex multi-objective optimization. The resulting H-infinity filter guarantees asymptotic stability of the estimation error dynamics with exponential convergence and is robust against nonlinear additive uncertainty and time-varying parametric uncertainties. Explicit bounds on the nonlinear uncertainty are derived based on norm-wise and element-wise robustness analysis.     

A new approach on designing l1 optimal regulator with minimum order for SISO linear systems

For a SISO linear discrete-time system with a specified input signal, a novel method to realize optimal l1 regulation control is presented. Utilizing the technique of converting a polynomial equation to its corresponding matrix equation, a linear programming problem to get an optimal l1 norm of the system output error map is developed which includes the first term and the last term of the map sequence in the objective function and the right vector of its constraint matrix equation, respectively. The adjustability for the width of the constraint matrix makes the trade-off between the order of the optimal regulator and the value of the minimum objective norm become possible, especially for achieving the optimal regulator with minimum order. By norm scaling rules for the constraint matrix equation, the optimal solution can be scaled directly or be obtained by solving a linear programming problem with l\ norm objective.     

A new approach on designing l_1 optimal regulator with minimum order for SISO linear systems

For a SISO linear discrete-time system with a specified input signal, a novel method to realize optimal l1 regulation control is presented. Utilizing the technique of converting a polynomial equation to its corresponding matrix equation, a linear programming problem to get an optimal l1 norm of the system output error map is developed which includes the first term and the last term of the map sequence in the objective function and the right vector of its constraint matrix equation, respectively. The adjustability for the width of the constraint matrix makes the trade-off between the order of the optimal regulator and the value of the minimum objective norm become possible, especially for achieving the optimal regulator with minimum order. By norm scaling rules for the constraint matrix equation, the optimal solution can be scaled directly or be obtained by solving a linear programming problem with l\ norm objective.     

非线性离散奇异摄动系统的组合优化控制方法的合理性分析

分析一类非线性离散奇异摄动系统的降阶组合优化控制器的合理性, 即降阶组合控制器与原始高阶优化控制器之间的关系. 基于快、慢子系统的解耦, 分别对快、慢子系统设计子优化控制器, 并进一步提出作用于原高阶系统的组合优化控制器. 对原高阶系统设计传统高阶优化控制器, 提出组合优化控制器近似等于传统高阶优化控制器的充分条件. 最后通过仿真验证了所得到结论的正确性.

     

A variable structure convex programming based control approach for a class of uncertain linear systems

A variable structure convex programming based control for a class of linear uncertain systems with accessible state is presented in this note. A convex programming problem is solved, on-line, by reformulating the problem in terms of a piecewise smooth penalty function, and relying on a suitable analog variable structure system implementing the gradient procedure. In this way, the controlled system reference movement, optimal with respect to a pre-specified scalar convex cost function and a set of suitable equality and inequality constraints, is generated. An inner control loop aimed at the finite time exact tracking of the reference movement is also designed. As a result, the controlled system trajectory starting in the feasible region there remains, and the optimal movement in the feasible region is proved to be an asymptotically stable equilibrium point of the controlled system.     

Optimal design of linear control systems by an interactive optimization method

When designing linear control systems, one of the most difficult problems is that the designer almost has no theoretical basis for the determination of proper parameters in order to obtain a system with desired specifications. Poles and directions of eigenvectors in the pole assignment method or weighting matrices of the quadratic criterion function in the optimal regulator method are such parameters. The designer has to determine them by trial-and-error using computer simulation. The purpose of this paper is to propose an approach to helping determine proper parameters in linear control system design by the state space methods. In the case where the desired specifications are not given explicitly, the approach applies an interactive optimization method called the Interactive Simplex method to search the most suitable parameters directly in the parameter space. But, if the specifications are given explicitly, the design problem can be formulated as a multiobjective optimization problem. In this case, weights which indicate relative importance of different specifications are introduced and the Interactive Simplex method is applied in the weight space to indirectly find the most appropriate parameters. The approach is implemented as part of a CAD system. The designer has only to make pairwise comparisons of response curves which are shown on a graphics display terminal in order to obtain the most preferred control system. Two illustrative examples are demonstrated to indicate the efficiency of the approach.     

Filter design: a finite dimensional convex optimization approach

We consider the design of transfer functions (filters) satisfying upper and lower bounds on the frequency response magnitude or on phase response, in the continuous and discrete time domains. The paper contribution is to prove that such problems are equivalent to finite dimensional convex optimization problems involving linear matrix inequality constraints. At now, such optimization problems can be efficiently solved. Note that this filter design problem is usually reduced to a semi infinite dimensional linear programming optimization problem under the additional assumption that the filter poles are fixed (for instance, when considering FIR design). Furthermore, the semi infinite dimensional optimization is practically solved, using a gridding approach on the frequency. In addition to be finite dimensional, our formulation allows to set or not the filter poles. These problems were mainly considered in signal processing. Our interest is to propose an approach dedicated to automatic control problems. In this paper, we focus on the following problems: design of weighting transfers for H_∞ control and design of lead‐lag networks for control. Numerical applications emphasize the interest of the proposed results. Copyright © 2003 John Wiley & Sons, Ltd.