Structure Simplification of Reduced Nonlinear Systems Using Efficient Fitness Functions and Search Spaces

Abstract

Known methods of model order reduction generally result in reduced systems with high number of internal interconnections, i.e. the model structures are complex. Starting from the system matrices optimization method, secondary conditions can be formulated to calculate reduced systems with simpler structures. One of the methods to find these secondary conditions is exploiting genetic algorithm in order to perform a global search within the search space. In this paper a modified fitness function is presented, which simplifies the search procedure and enormously reduces the computation effort. Also a method for omitting improper candidates is suggested that accelerates the whole search process by shrinking the search space.