Integrated optimization based on successive adaptive modeling

In order to meet the high requirement of practical optimization, it is essential to construct an integrated optimization system for real systems, which is a new paradigm of optimization system based on combining optimization technique with modeling and simulation technologies. From the viewpoint of optimality and computational efficiency, this paper examines some strategies for arranging sample points adaptively in an integrated optimization system that combines Particle Swarm Optimization and Radial Basis Function Network. The proposed strategies for arranging sample points are examined through numerical simulations using four types of typical benchmark problems.