Benchmarking State-of-the-art DIRECT-type Methods on the BBOB Noiseless Testbed

Abstract

In recent years, there has been significant progress in the development of new DIRECT-type algorithms for black-box optimization problems. In this paper, we evaluate three well-performing DIRECT-type methods from a recent extensive numerical study on the BBOB noiseless testbed in dimensions 2, 3, 5, 10, and 20. We discuss the strengths and weaknesses of these algorithms on different classes of functions and provide a comparison with the original DIRECT method, as well as with three other well-established methods: RL-SHADE, L-BFGS-B, and SLSQP.

In recent years, there has been significant progress in the development of new DIRECT-type algorithms for black-box optimization problems. In this paper, we evaluate three well-performing DIRECT-type methods from a recent extensive numerical study on the BBOB noiseless testbed in dimensions 2, 3, 5, 10, and 20. We discuss the strengths and weaknesses of these algorithms on different classes of functions and provide a comparison with the original DIRECT method, as well as with three other well-established methods: RL-SHADE, L-BFGS-B, and SLSQP.