MetaBox

MetaBox是由华南理工大学开发的一个用于元黑盒优化的基准平台，专门为开发和评估元黑盒优化与强化学习（MetaBBO-RL）方法设计。该平台提供了一个包含300个问题实例的广泛测试集，涵盖从合成到真实场景的多样化问题。MetaBox还集成了19种基准方法，包括传统的黑盒优化器和最新的MetaBBO-RL方法。此外，MetaBox引入了三种标准化的性能指标，以更全面地评估方法的效果。该平台旨在简化算法开发，提供自动化的工作流程，并支持对MetaBBO-RL方法的深入分析和严格评估。

MetaBox is a benchmark platform for meta black-box optimization developed by South China University of Technology, specifically designed for developing and evaluating meta black-box optimization and reinforcement learning (MetaBBO-RL) methods. The platform provides a comprehensive test suite containing 300 problem instances, covering diverse scenarios ranging from synthetic to real-world applications. Additionally, MetaBox integrates 19 benchmark methods, including traditional black-box optimizers and state-of-the-art MetaBBO-RL approaches. Furthermore, MetaBox introduces three standardized performance metrics to enable more comprehensive evaluation of method effectiveness. This platform aims to simplify algorithm development, offer automated workflows, and support in-depth analysis and rigorous evaluation of MetaBBO-RL methods.