A Preference-Based Physical Programming Method for Multi-Objective Designs of Electromagnetic Devices

Abstract

The optimal algorithm plays a crucial role in optimizations of electromagnetic devices in terms of both solution efficiency and precision of the final result. To reduce the unnecessary information bombing of a decision maker (DM) and to improve the convergence ability of solution procedures for multi-objective design problems, a vector preference-based physical programming method is proposed. To specify the region of interests of a DM, a preference frame is constructed using an aspiration point, a reservation point, and a preference vector; to make a full use of the explored information, the sampling points are classified based on all the pseudo preferences in one single run, and the aggregated preference functions are calculated in parallel. A dynamic pseudo preference translation offset vector is developed to accommodate different mapping mechanisms for a variety of diverse optimization problems and increase the robustness of the algorithm. Typical test functions and inverse problems are solved to demonstrate the effectiveness and efficiency of the proposed method.