Cooperative parking search: Reducing travel time by information exchange among searching vehicles

In dense inner-city regions it is often difficult to find a free parking space. Vehicles spend significant time searching for a parking space near their intended destination and, as a result, city traffic grows. With new technologies, it is possible for searching vehicles to share parking-related information to increase search efficiency. In this paper, we investigated whether cooperation during parking search reduces search time as well as total travel time, which includes the final stage of walking to the destination. We modeled the cooperation between vehicles using a centralized approach: vehicles repeatedly send their current location, destination, and their past and intended search routes to a central instance. This instance optimizes the search routes and sends back an individualized route suggestion to each vehicle. Using the traffic simulator SUMO, we designed an exemplary city district in which vehicles search for roadside parking spaces. We divided the active search process into two stages: before reaching the destination and after reaching the destination without having found a parking space yet. For the first stage, we developed an algorithm which incorporates a modification of Dijkstra's shortest-path algorithm to reduce “double-searching” of street segments. If no parking space has been found prior to reaching the destination, an iterative decision on the next road segment based on a cost function is applied. The computational results showed benefits in search time for all investigated cases with a reduction up to 30 percent, when many vehicles search and parking spaces are limited. Search time reductions were observed both when approaching the destination and when circling after passing the destination.