Toward Automated Algorithm Design: A Survey and Practical Guide to Meta-Black-Box-Optimization

TL;DR

MetaBBO proposes a bi-level, data-driven framework for automated black-box optimization algorithm design, unifying four meta tasks (AS, AC, SM, AG) under four learning paradigms (RL, SL, NE, ICL). Across a comprehensive survey and empirical evaluation on the MetaBox benchmark, it shows that RL-based MetaBBO methods can achieve competitive optimization performance and generalization, while SM and AG offer end-to-end and creative algorithm design with notable computational costs. The work emphasizes crucial design choices in neural architectures, state representations, and benchmark construction to improve generalization, and outlines future directions toward task mixtures, fully autonomous pipelines, and smarter LLM integration. By providing a structured taxonomy, practical guidance, and benchmarking insights, this survey aims to accelerate the development and application of automated algorithm design in black-box optimization.

Abstract

In this survey, we introduce Meta-Black-Box-Optimization~(MetaBBO) as an emerging avenue within the Evolutionary Computation~(EC) community, which incorporates Meta-learning approaches to assist automated algorithm design. Despite the success of MetaBBO, the current literature provides insufficient summaries of its key aspects and lacks practical guidance for implementation. To bridge this gap, we offer a comprehensive review of recent advances in MetaBBO, providing an in-depth examination of its key developments. We begin with a unified definition of the MetaBBO paradigm, followed by a systematic taxonomy of various algorithm design tasks, including algorithm selection, algorithm configuration, solution manipulation, and algorithm generation. Further, we conceptually summarize different learning methodologies behind current MetaBBO works, including reinforcement learning, supervised learning, neuroevolution, and in-context learning with Large Language Models. A comprehensive evaluation of the latest representative MetaBBO methods is then carried out, alongside an experimental analysis of their optimization performance, computational efficiency, and generalization ability. Based on the evaluation results, we meticulously identify a set of core designs that enhance the generalization and learning effectiveness of MetaBBO. Finally, we outline the vision for the field by providing insight into the latest trends and potential future directions. Relevant literature will be continuously collected and updated at https://github.com/MetaEvo/Awesome-MetaBBO.

Figures (10)

• Figure 1: Roadmap of the content structure, beginning with a concept introduction, followed by a review of existing methods across different taxonomies, a evaluation of selected methods, and a summary of key design strategies and future vision.
• Figure 2: A conceptual overview of the bi-level learning framework of MetaBBO, illustrating the interactions between its core components to clarify the overall workflow.
• Figure 3: Conceptual workflow of MetaBBO for Algorithm Selection.
• Figure 4: Conceptual workflow of MetaBBO for Algorithm Configuration.
• Figure 5: Conceptual workflow of MetaBBO for Solution Manipulation.