Dynamical inverse problems of distributed systems (Q1413067)

This monograph develops methods for reconstructing the input (a control input, a disturbance input and/or an initial condition) from the measured output of a system described by an evolution equation. Because there is an imprecision in the measured output, there is an ambiguity in the solution and the problem is ill-posed. One hopes that the solution(s) for a given level of imprecision in the measurements will tend to the true solution when the imprecision tends to zero. In case there are many solutions for a certain imprecision, an optimization problem on the corresponding set of inputs can be treated. The resulting optimal input should tend to the true solution when the imprecision decreases. This dynamic situation generalizes the static one corresponding to identification of distributed systems. The specificity here is that one looks for feedback type inputs for certain kinds of distributed systems whith associated models. In chapter one, the abstract case is addressed. In chapter two, one studies parabolic equations (distributed or boundary controls). Chapter three is devoted to parabolic variational inequalities. Chapter four follows a certain tradition of proposing computational methods for a developed theory. Notice that in the first pages the author defines his objects on sets and then he introduces functionals on those sets. This is quite general for doing analysis, and the Banach space case comes soon. A bibliography can be found at the end, but there is no index. The book is of interest for those studying inverse and optimization problems related to distributed systems, in particular with respect to input determination and identification.