Linear Systems Control Solution

Symmetries and Semi-invariants in the Analysis of Nonlinear Systems details the analysis of continuous- and discrete-time dynamical systems described by differential and difference equations respectively.

In this chapter we discuss linear ordinary differentia! equations. For the 1110st part we will not find it necessary to assume that these equation are also time invariant, although this is an important special case. We are interested in exposing the structure of the solutions and, to a lesser extent, developing actual solution techniques.

HIS book presents a generalized state-space theory for the analysis T and synthesis of finite horizon suboptimal Hoo controllers. We de­ rive expressions for a suboptimal controller in a general setting and propose an approximate solution to the Hoo performance robustness problem. The material in the book is taken from a collection of research papers written by the author. The book is organized as follows. Chapter 1 treats nonlinear optimal control problems in which the cost functional is of the form of a quotient or a product of powers of definite integrals. The problems considered in Chap­ ter 1 are very general, and the results are useful for the computation of the actual performance of an Hoo suboptimal controller. Such an application is given in Chapters 4 and 5. Chapter 2 gives a criterion for the evaluation of the infimal Hoc norm in the finite horizon case. Also, a differential equation is derived for the achievable performance as the final time is varied. A general suboptimal control problem is then posed, and an expression for a subopti­ mal Hoo state feedback controller is derived. Chapter 3 develops expressions for a suboptimal Hoo output feedback controller in a very general case via the solution of two dynamic Riccati equations. Assuming the adequacy of linear expressions, Chapter 4 gives an iterative procedure for the synthesis of a suboptimal Hoo controller that yields the required performance even under parameter variations.

A unified approach is proposed for applied mechanics and optimal control theory. The Hamilton system methodology in analytical mechanics is used for eigenvalue problems, vibration theory, gyroscopic systems, structural mechanics, wave-guide, LQ control, Kalman filter, robust control etc. All aspects are described in the same unified methodology. Numerical methods for all these problems are provided and given in meta-language, which can be implemented easily on the computer. Precise integration methods both for initial value problems and for two-point boundary value problems are proposed, which result in the numerical solutions of computer precision.