Probabilistic Routing for On-Street Parking Search

摘要: An estimated 30% of urban traffic is caused by search for parking spots [Shoup, 2005]. Suggesting routes along highly probable could reduce traffic. In this paper, we formalize as a probabilistic problem on road graph and show that it NP-complete. We explore heuristics optimize the driving duration walking distance to destination. Routes are constrained reach certain probability threshold finding spot. Empirically probabilities successful attempts provided TomTom per-street basis. release these dataset about 80,000 roads covering Berlin area. This allows evaluate algorithms real network with realistic first time. However, many other areas, not openly available. Because they effortful collect, propose an algorithm relies conventional attributes only. Our experiments comes close baseline factor 1.3 in our cost measure. leads conclusion may be sufficient compute reasonably good routes.