In this paper it is considered the possibility to utilize models of nonlinear systems obtained by evolving radial base network for solving multi-criteria optimization task. The solution to this problem is to find the Pareto-optimal set of solutions, which involves multiple simulation of optimized system for different values of variable parameters. In real systems it is usually requires significant computational costs what raises the need to construct their high-speed models. Simulation modeling showed that quite accurate models may be received by evolving radial basic network, what allow determining the set of Pareto-optimal solutions.
multiobjective optimization, neural network, genetic algorithm, nonlinear system, Pareto set, optimal solution
"Mnohokryteryalnaia neiroэvoliutsyonnaia optymyzatsyia nelyneinыkh funktsyi" ,
Information Processing Systems,