Search results

This paper aims to propose a numerically efficient multi-objective optimization strategy, which can improve both the efficiency and performance during the optimization process.

This paper discusses the multi-objective optimization algorithm by combining multi-objective differential evolution (MODE) algorithm with an adaptive dynamic Taylor Kriging (ADTK) model.

The proposed approach is validated through application to an analytic example and applied to a shape optimal design of a multi-layered interior permanent magnet synchronous motor for torque ripple reduction while maintaining the average torque.

The ADTK model selects its basis functions adaptively and dynamically so that it may have better accuracy than any other Kriging models. Through adaptive insertion of new sampling data, it guarantees minimum required sampling data for a desired fitting accuracy.