Black-Box Optimization Revisited: Improving Algorithm Selection Wizards through Massive Benchmarking

TL;DR

This paper introduces a comprehensive benchmark suite, OptimSuite, and an automated algorithm selection tool, ABBO, which together improve generalization and performance in black-box optimization tasks across diverse problem types.

Contribution

The work presents a broad benchmark collection and a general-purpose algorithm selection wizard that outperforms previous methods without task-specific tuning.

Findings

ABBO achieves competitive results across diverse benchmarks.

ABBO significantly outperforms previous state-of-the-art algorithms.

The benchmark suite covers a wide range of real-world and academic problems.

Abstract

Existing studies in black-box optimization for machine learning suffer from low generalizability, caused by a typically selective choice of problem instances used for training and testing different optimization algorithms. Among other issues, this practice promotes overfitting and poor-performing user guidelines. To address this shortcoming, we propose in this work a benchmark suite, OptimSuite, which covers a broad range of black-box optimization problems, ranging from academic benchmarks to real-world applications, from discrete over numerical to mixed-integer problems, from small to very large-scale problems, from noisy over dynamic to static problems, etc. We demonstrate the advantages of such a broad collection by deriving from it Automated Black Box Optimizer (ABBO), a general-purpose algorithm selection wizard. Using three different types of algorithm selection techniques, ABBO achieves competitive performance on all benchmark suites. It significantly outperforms previous state of the art on some of them, including YABBOB and LSGO. ABBO relies on many high-quality base components. Its excellent performance is obtained without any task-specific parametrization. The OptimSuite benchmark collection, the ABBO wizard and its base solvers have all been merged into the open-source Nevergrad platform, where they are available for reproducible research.