Dynamic pricing for improved parking space utilization: a data-driven Bayesian optimization approach

Parking scarcity in downtown areas presents a major challenge for commuters seeking convenient spaces near their destinations. This study proposes a dynamic pricing optimization framework for garage parking in downtown Chicago, balancing the often-conflicting objectives of garage operators and users. The framework formulates garage pricing as a multi-objective optimization problem that combines normalized garage revenue and occupancy, with the weight factors calibrated through a sensitivity analysis to identify the trade-off that maximizes occupancy gains while approximately preserving baseline garage revenue. Using simulated POLARIS data, we applied a Bayesian optimization framework to determine the optimal garage prices for both peak and off-peak periods, then evaluated key performance metrics, including the temporal and spatial distribution of garage fees, occupancy rates, and revenues over a 24-hour period. Results showed that the optimized pricing strategy substantially reduced off-peak garage fees, improving accessibility during off-peak periods, while peak-period fees remained close to existing baseline levels, avoiding disadvantage to peak-period travelers. Peak-hour occupancy increased from approximately 50% to over 70%, with improved overall parking space utilization and limited revenue trade-off. The study also highlights the influence of allocating a portion of parking spaces to other road users.