Extended Multipoint Approximation Method

Abstract

Stemming from polynomial metamodels, multipoint approximation method (MAM) and moving least square method (MLSM) focus on the development of metamodels for the objective and constraint functions in solving a mid-range optimization problem with a trust region. Although both of these methods could solve problems successfully, there is still some room for improvement on the computational effort and search capability. To address this problem, the extended multipoint approximation method is proposed to seek the optimal solution in this paper. The developed method assimilating the advantage of Taylor’s expansion used in MLSM demonstrates its superiority over other methods in terms of the computational efficiency and accuracy by some well-established benchmark problems.