{"id":"arxiv/2604.15492","url":"https://arxiv.org/abs/2604.15492","tags":["math.OC"],"score":0,"title":"Enhancing Model Based Derivative Free Optimization using Direct Search","domain":"math.OC","source":"arxiv","authors":["Zijun Li","Aswin Kannan"],"reviews":0,"subject":"math.OC","summary":"We consider single and multiobjective simulation-based optimization problems. Simulation-based optimization has traditionally used both model-based and search-based methods, often in isolation. Model-based methods include trust region approaches and Bayesian optimization, while search methods include genetic algorithms and Direct Search-type techniques. In this work, we propose a switching framework that leverages Direct Search methods to enhance the performance of any model-based optimizer. Our contributions are twofold. First, in the single-objective setting, we analyze and prove the…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-04-16","reactions":{}}