A Surrogate-Assisted Cooperative Co-evolutionary Algorithm for Solving High Dimensional, Expensive and Black Box Optimization Problems

Many research efforts have been recently focus to solve large-scale global optimization (LSGO) problems by means of evolutionaryalgorithms. Cooperative co-evolution has been proposed to solve suchproblems depending on thousands of variables. This methodology hasproved very efficient in solving a wide range of LSGO problems. Never-theless, it often requires an extremely large number of function evalua-tions to reach a suitable solution. This is somewhat problematic whenthe function evaluation is computationally expensive. A globally effectiveapproach to high-fidelity optimization problems based on such expensiveanalyses lies in the exploitation of surrogate models. They act as cheap-to-evaluate alternatives to the original high-fidelity models reducing thecomputational cost, while still providing improved designs. This kind ofoptimization process, referred to as surrogate-assisted optimization, hasproved very efficient on small-dimensional problems but suffers from thecurse of dimensionality to solve LSGO problems. In this paper, coop-erative co-evolution was combined with surrogate-assisted optimizationin order to efficiently solve high dimensional, expensive and black-boxproblems. Experimental results are provided on a wide set of benchmarkproblems and show promising results for the proposed algorithm.