Proactive pricing strategies for on-street parking management with physics-informed neural networks

Abstract

Effective pricing is important for on-street parking management and proactive parking pricing is an innovative strategy to achieve optimal parking utilization. For proactive parking pricing, accurately predicting parking occupancy and deriving the price elasticity of parking demand are necessary. In recent years, there have been an increasing number of studies applying big data technology for parking-occupancy prediction. However, existing research has not incorporated economic knowledge into modeling, thus preventing application of the price elasticity of parking demand. In this study, proactive pricing strategies are proposed to adjust on-street parking prices which involve a parking-occupancy prediction model and a price-optimization method. Physics-informed neural networks are employed to achieve accurate prediction of parking occupancy and calculation of parking price elasticity. An elasticity-occupancy parking-management strategy is proposed for on-street parking management which leverages parking occupancy and price elasticity to guide pricing interventions. A case study shows that the parking-occupancy prediction model can make accurate predictions and derive the price elasticity of parking demand. Proactive parking pricing enables drivers to plan their trips in advance, allowing parking occupancy within an optimal range.