Model reduction of discrete interval systems using genetic algorithms

In this paper, an evolutionary approach is proposed to derive a reduced-order model for discretetime interval systems based on resemblance of discrete sequence energy between the original and reduced systems. With the use of the recursive algebraic algorithm and interval arithmetic manipulations, the problem to identify boundaries of the uncertain coefficients of the reduced-order model can be formulated as an optimization problem, which is subsequently solved by a proposed genetic algorithm. To demonstrate the effectiveness of the proposed approach, system performance of the reduced-order discrete interval model is validated based on time responses in comparison to existing approaches. Because of the time-consuming process that genetic algorithms generally exhibit, particularly the problem nature which demands heavy calculation of the fitness function, a parallel computation scheme is also presented to accelerate the evolution process to derive the reduced-order model.