Multi-Criteria Modeling for Smart Parking and Parking Reservation

Abstract

In the realm of smart cities, issues related to vehicle parking have increasingly contributed to traffic congestion, primarily due to drivers searching for vacant spots and the inefficient management by parking operators. Therefore, smart parking systems must continuously be enhanced with real-time models that can reflect the availability of parking spaces and improve the utilization of parking capacity, thereby addressing part of these challenges. This paper proposes a two-stage hybrid approach aimed at optimizing parking space reservations in smart parking systems. The model enables drivers to locate the most time-efficient parking spot within these systems. In the first stage, the parking space is evaluated within an environment defined by fixed capacity. In the second stage, the most suitable parking spot is made available through a hierarchical parking management system using a search-based method. The approach is grounded in Markov Chain modeling, which is employed to accurately estimate the number of vehicles requiring reservations. Underestimation of this figure results in insufficient demand adjustments, whereas overestimation leads to additional operational costs. Therefore, the proposed system uses Markov Chains for capacity forecasting and mathematical models to simulate the temporal movement of vehicles within the parking facility to determine the proportion of reserved versus regular parking spaces. This temporal analysis improves reservation predictions and yields an average occupancy rate of approximately 0.032, with 15 successful reservations recorded over a 24-hour period.