Algorithm Selection

Algorithm selection aims to automatically choose the best algorithm for a given problem instance, improving efficiency and performance compared to using a single, general-purpose algorithm. Current research focuses on developing machine learning models, including neural networks (e.g., deep recurrent networks, convolutional neural networks, transformers) and tree-based methods (e.g., Random Forests, XGBoost), to predict optimal algorithm choices based on problem features or algorithm trajectories. This field is significant because it can enhance the performance of various applications, from optimization problems in operations research and machine learning to image processing and multi-agent pathfinding, by leveraging the complementary strengths of different algorithms.