Series On Optimization - Böcker

This monograph deals with the controlled/non-controlled nonlinear systems of differential equations. A mathematical apparatus is developed to construct stationary conditions and to carry out studies on the behaviour of integral curves in the neighbourhood of such conditions.Considerable coverage is given to existence and methods of finding periodic orbits and almost-periodic solutions, as well as to the description of the class of ergodic recurrent motions. There is further treatment of the perturbation method and the theory of time-independent and periodic perturbations in particular.The theory developed here is applied to the construction and investigation of the neigbourhood of time-independent conditions for nonlinear systems of automatic control, and the control of charged particle beam in magnetic field. Some other specific problems are also solved such as after effect systems and orbit quantization.

The classical optimal control theory deals with the determination of an optimal control that optimizes the criterion subjects to the dynamic constraint expressing the evolution of the system state under the influence of control variables. If this is extended to the case of multiple controllers (also called players) with different and sometimes conflicting optimization criteria (payoff function) it is possible to begin to explore differential games. Zero-sum differential games, also called differential games of pursuit, constitute the most developed part of differential games and are rigorously investigated. In this book, the full theory of differential games of pursuit with complete and partial information is developed. Numerous concrete pursuit-evasion games are solved (”life-line” games, simple pursuit games, etc.), and new time-consistent optimality principles in the n-person differential game theory are introduced and investigated.

This monograph defines the notion of a “system” by reference to those systems which exhibit goal-oriented behavior and utilize the notion of decision making and controls. Such systems allow for phenomenological description and fix the nature of causal transformations of input effects into output quantities. The study of consequences of the fact that the systems possess some properties constitutes the content of systems optimization methodology which goes beyond the scope of descriptive classification of systems.Chapter 1 deals with philosophical problems of systems methodology. An attempt is made to systematize and analyze the problems of scientific methodology as applied to systems modeling methodology which is viewed as the most general concept utilized in modern science.Chapter 2 focuses on problems of qualitative analysis in natural and social sciences. Attention is drawn to problems of measurement theory and quantitative analysis of systems.Approaches and methods of systems analysis and synthesis form the central portion of the book. Much study is given to the methods of systems decomposition, an integration using both discrete and continuous descriptions of objects, processes, and phenomena. Examples of complex goal-oriented systems are also provided.The remaining part of the book is largely centered around the methodology of multiobjective systems optimization.