An Advanced and Adaptive Tabu Search Algorithm for Dynamic Shared Parking Reservation and Allocation

Inefficient utilization of existing parking resources is the main cause of parking difficulties, especially in metropolises. Shared parking allocation based on the online parking reservation system (PRS) is an effective way to deal with the inefficient utilization since it increases available parking resources through sharing and avoids blind search through reservation and allocation. This study focuses on the dynamic shared parking allocation problem based on PRS. A metaheuristic algorithm, namely Advanced and Adaptive Tabu Search (AATS), was designed to cope with the real-time updates of parking demands and shared parking space and to achieve good allocation effect and high allocation speed. The proposed algorithm is based on advanced initialization with multi-factor sequencing and on adaptive neighborhood generation with bi-operator competition. The results of a three-day district-level experiment show that AATS achieves an allocation effect close to the exact algorithm, while having a significant superiority in allocation speed. The proposed AATS is practical to deal with realistic parking problems.