Optimizing Control System Based on Integration of Competing Search Optimization Algorithms

摘要

The paper considers an approach to development of an optimizing control system (OCS) based on the combined use of competing algorithms of search for fitness function optimum. The approach implies application of an algorithm to resolve contradictions between different search algorithms. This algorithm has a fundamental difference from well-known typical approaches to resolving contradictions that either involve averaging a control signal or selection of one of them based on estimation of the control system status or the environment it operates in. In some engineering cases, this provides simultaneous advantages for each separate algorithm. A distinctive feature of every single algorithm is that it may ensure a high efficiency under certain conditions, the developed setting methodology, and use of effective methods of estimation of future, current or past states of the system. The paper reveals possible setup procedures for the proposed system implying quality operation based on separate search algorithms at the first stages followed by adjustment of parameters. Examples of search processes under conditions of drifting of the extremum static characteristic are stated.