Abstract
Meta-heuristic algorithms, which require a large number of fitness evaluations before locating the global optimum, are often prevented from being applied to computationally expensive real-world problems where one fitness evaluation may take from minutes to hours, or even days. Although many surrogate-assisted meta-heuristic optimization algorithms have been proposed, most of them were developed for solving expensive problems up to 30 dimensions. In this paper, we propose a surrogate-assisted hierarchical particle swarm optimizer for high-dimensional problems consisting of a standard particle swarm optimization (PSO) algorithm and a social learning particle swarm optimization algorithm (SL-PSO), where the PSO and SL-PSO work together to explore and exploit the search space, and simultaneously enhance the global and local performance of the surrogate model. Our experimental results on seven benchmark functions of dimensions 30, 50 and 100 demonstrate that the proposed method is competitive compared with the state-of-the-art algorithms under a limited computational budget.