On Traveling Salesperson Problems for Dubins ’ vehicle: stochastic and dynamic environments

In this paper we propose some novel planning and routing strategies for Dubins’ vehicle, i.e., for a nonholonomic vehicle moving along paths with bounded curvature, without reversing direction. First, we study a stochastic version of the Traveling Salesperson Problem (TSP): given n targets randomly sampled from a uniform distribution in a rectangle, what is the shortest Dubins\´tour through the targets and what is its length? We show that the expected length of such a tour is Ω(n^m=2/3)and we propose a novel algorithm that generates a tour of length O(n^m=2/3log(n)^m=1/3) with high probability. Second, we study a dynamic version of the TSP (known as "Dynamic Traveling Repairperson Problem" in the Operations Research literature): given a stochastic process that generates targets, is there a policy that allows a Dubins vehicle to stabilize the system, in the sense that the number of unvisited targets does not diverge over time? If such policies exist, what is the minimum expected waiting period between the time a target is generated and the time it is visited? We propose a novel receding-horizon algorithm whose performance is almost within a constant factor from the optimum.