论文标题
混合参数搜索和混合变量贝叶斯优化的动态模型选择
Hybrid Parameter Search and Dynamic Model Selection for Mixed-Variable Bayesian Optimization
论文作者
论文摘要
本文提出了一种用于管理混合变量的贝叶斯优化（BO）的新型混合模型，包括定量（连续和整数）以及定性（分类）类型。我们提出的新混合模型（命名为Hybridm）合并了蒙特卡洛树搜索结构（MCTS），用于与连续的高斯工艺（GP）的分类变量合并。 Hybridm利用上置信度绑定的树搜索（UCT）进行MCTS策略，从而展示了树架构的集成到贝叶斯优化中。我们的创新，包括在替代建模阶段中的动态在线内核选择以及独特的UCT搜索策略，将混合模型定位为混合可变替代模型的进步。数值实验强调了混合模型的优越性，突出了它们在贝叶斯优化中的潜力。
This paper presents a new type of hybrid model for Bayesian optimization (BO) adept at managing mixed variables, encompassing both quantitative (continuous and integer) and qualitative (categorical) types. Our proposed new hybrid models (named hybridM) merge the Monte Carlo Tree Search structure (MCTS) for categorical variables with Gaussian Processes (GP) for continuous ones. hybridM leverages the upper confidence bound tree search (UCTS) for MCTS strategy, showcasing the tree architecture's integration into Bayesian optimization. Our innovations, including dynamic online kernel selection in the surrogate modeling phase and a unique UCTS search strategy, position our hybrid models as an advancement in mixed-variable surrogate models. Numerical experiments underscore the superiority of hybrid models, highlighting their potential in Bayesian optimization.
