Advances in meta-heuristic methods for large-scale black-box optimization problems

Abstract

The optimal design of complex engineering equipment usually faces high-complexity, high-dimensional optimization problems – the so-called "large-scale black-box optimization problems (LBOPs)" – which are characterized by unavailable mathematical expressions of objective functions and/or constraint functions, and high dimensionality of design variables. The LBOPs have attracted the interest of scholars in various fields in recent years, and meta-heuristic algorithms are considered effective methods for solving these problems. This paper comprehensively summarizes recent research progress in meta-heuristic algorithms for solving LBOPs, including meta-heuristic algorithms with and without decomposition strategies, and meta-heuristic algorithms for handling computationally expensive large-scale optimization problems. Finally, possible future research directions of meta-heuristic methods for solving LBOPs are proposed.

Keywords

• large-scale optimization
• black-box optimization
• meta-heuristic algorithms
• surrogate-assisted optimization