Output details

A Gaussian Process Surrogate Model Assisted Evolutionary Algorithm for Medium Scale Expensive Optimization Problems

<22> This paper focuses on surrogate model assisted evolutionary algorithms (SAEAs) for medium-scale (20-50 decision variables) computationally expensive optimisation problems, which have not been well studied yet. Two fundamental contributions have been presented: (1) A surrogate model-aware search framework was proposed. Better or similar solutions can be obtained with 12% to 50% exact function evaluations compared to several state-of-the-art SAEA frameworks. (2) Dimension reduction techniques were introduced into SAEA research for tackling the “curse of dimensionality”. Based on this research, the first practical method for mm-wave IC synthesis and the first efficient and scalable antenna design optimisation methods have been proposed.