Dynamic parking space allocation at urban scale: Problem formulation and resolution

Abstract

Searching for an available parking space is a major cause of traffic congestion and pollution. Besides, getting a free spot constitutes a primary worry for drivers on the road, which results in their daily discomfort and stress. Therefore, well-designed parking management tools are needed to effectively allocate parking spaces to drivers, especially in dense urban areas. In this perspective, we propose two linear integer programming models for assigning parking spaces to all road users (private vehicles and carriers) while maximizing parking occupancy. The first combinatorial problem attributes to each driver only one parking space while the second one assigns two parking spaces if no space is suitable for that motorist. Feasible solutions are computed in an exact way by improved primal simplex method and in an approximated manner by two metaheuristics (Genetic algorithm and Taboo search) for larger scale applications. However, parking problems still persist when parking requests are not properly balanced between a city's zones, a local gridlock problem arises where certain areas are over-used and others under-used. For this reason, we first redistribute parking demands over urban areas before dedicating one or two spaces to each driver. To do this, we consider both solicitation rates in these areas and coverage ratios of received parking demands. Our parking space allocation solution is just one component of a transportation system heavily backed by information and communication technologies. This system consists of effectively managing a city's parking spaces by predicting and understanding their occupancy patterns. The suggested approach is applied to Casablanca city with several parking scenarios in order to evaluate its effectiveness and test its ability to minimize total walking distance, parking costs as well as unmet demand percentages. Experimental results show that our mathematical models can effectively rule out parking problems. The related published methods satisfy at most 75% of received parking requests. However, with our proposal, it is possible to find parking spaces for 87.3% of drivers without any noticeable effort, even an increase of 16.4%. The ordinary walking distance is reduced by 40% if our solution is adopted. Yet, other research works report a decrease of this distance not exceeding 30%. The increase in parking demand satisfaction rates demonstrates the effectiveness of our system, especially in the eyes of urban stakeholders. The reduction in walking distance encourages drivers to use our solution and consequently increases their loyalty.

Keywords

• Freight transport
• Smart system
• Parking space allocation
• Integer linear programming
• Combinatorial problem
• Optimization algorithms