Reducing Street Parking Search Time via Smart Assignment Strategies

In dense metropolitan areas, searching for street parking adds to traffic congestion. Phone-based parking assistants have been proposed as a remedy, but evidence on their impact is limited. We quantify how coordination and information in such apps affect search time and the probability of finding on-street parking. Using a city-scale, data-driven simulation of Madrid, we compare four strategies for app-enabled drivers (participants) competing with other drivers (competitors): uncoordinated (Unc-Agn), coordinated without competitors' awareness (Cord-Agn), an oracle with full competitors' knowledge (Cord-Oracle), and our novel/practical method, Cord-Approx, that models competitors probabilistically. Rather than requiring knowledge of competitors' distances to each spot, Cord-Approx uses learned occupancies to inflate effective distances to alternatives, then dispatches via the Hungarian algorithm. In high-fidelity simulations of Madrid's parking network with real traffic data, Cord-Approx users averaged 6.69 minutes to park, versus 19.98 for competitors without an app. Across zones, Cord-Approx reduces the average search time by 76% in Culture & Transport Hubs, and 72% in Residential & Light Industry, relative to competitors.