The purpose of this book is to present some new algorithms for computing optimal controls for linear, multivariable systems. Time, fuel, and effort optimal problems with effort constraints are considered. The Maximum Principle is used extensively and some new results are reported both on normality and also for the case where the problem contains integral constraints on the control.
The algorithms can be extremely useful to the practicing engineer even if he has no desire to build an “optimal” system: a solution to the optimal control problem will allow him to make a better choice among new designs when a comparison to optimal performance, based on several performance criteria, can be made. Experimental verification is given for the theoretical conclusions presented, and it is shown that these algorithms can be of value to the designer of suboptimal control systems as well.
The author proceeds from a survey of known techniques to a mathematical analysis of his proposed procedures, proofs of convergence, and a description of experimental results. Fortran programs for realizing the algorithms are also given, and the book suggests several ways in which iterative procedures can be used in design. Of particular interest are the minimum fuel experimental results which may be the first reported for high-order systems. The minimum fuel solutions demonstrate that the time-optimal performance is unsatisfactory, since the fuel used could be greatly reduced with a small increase in the control interval.
Some Iterative Solutions in Optimal Control should interest all engineers and designers working with optimal control systems. The Fortran programs in particular can be used by the designer to compare the performance of a non-optimal designer with that of the optimal.