Warm-Starting Nested Rollout Policy Adaptation with Optimal Stopping

Nested Rollout Policy Adaptation (NRPA) is an approach using online learning policies in a nested structure. It has achieved great result variety of difficult combinatorial optimization problems. In this paper, we propose Meta-NRPA, which combines optimal stopping theory with NRPA for warm-starting and significantly improves the performance NRPA. We also present several exploratory techniques enable it to perform better exploration. establish three notoriously problems ranging from telecommunication, transportation coding namely Minimum Congestion Shortest Path Routing, Traveling Salesman Problem Time Windows Snake-in-the-Box. improve lower bounds Snake-in-the-Box problem multiple dimensions.