Computational methods for optimizing distributed systems (Q800866)

This six-chapter landmark book blending original, constructive, theoretical and computational exposition of optimal control theory of distributed parameter systems devises some computational algorithms based on the formalism of a first-order strong variational method or gradient- type methods for solving this class of problems. The prerequisites for the book are elements of measure theory, partial differential equations and a sound knowledge of functional analysis to a level of mathematical sophistication equivalent to that possessed within senior undergraduate mathematics courses. Optimal control of distributed parameter systems has been of great interest. In fact, many industrial processes using chemical reactors, furnaces, distillation columns, for examples are best modelled by partial-integro differential equations which are commonly called ''distributed parameter systems'' in engineering parlance. Chapter I expounds a simplified and unifying framework for an intimate understanding of the sequel by providing an array of pertinent definitions, fundamentals and notations adopted throughout the text. Chapter II summarises the basic theory of first and second boundary-value problems of linear second-order parabolic partial differential equations in both general and divergence forms carefully tailored towards forming a suitable basis for exploring its new applications in the later chapters. Chapter III, using a strong variational technique to devise a computational algorithm, rapidly deals with an optimal control problem involving first boundary-value problems of linear second-order parabolic partial differential equation corresponding to a class of stochastic optimal control problems equipped with Markov terminal term. Chapter IV extends the results of chapter III to the situation when both controls and parameter vectors are assumed to act on the coefficients of the differential operator associated with the problem under investigation. Chapter V presents and illustrates an elegant convergence theory for the algorithm developed in chapter III. Chapter VI is concerned with optimal control problems (necessary and sufficient conditions for optimality) involving second boundary-value problems of a linear second-order parabolic partial differential equation. And in order not to interrupt the continuity of the materials within the main text, the authors introduce a number of useful appendices which are appropriately referred to in the text.