An Adaptive Method for the Stochastic Orienteering Problem

We consider the NP-hard Stochastic Orienteering Problem, where goal is to navigate between start and end vertices in a graph, maximizing sum of rewards for visited while obeying travel budget over edges with stochastic cost within given probability failure. Previously, we solved this by finding an initial path using deterministic orienteering solver transformed it into policy Constrained Markov Decision Process that can skip based on arrival time. In work augment our technique, creating tree which branches at vertex-time states high being skipped, allowing new sequences resulting policy. demonstrate adaptive method collects significantly more reward expectation, even when number limited control computation