학술논문

Optimization Model of Autonomous Vehicle Parking Facilities, Developed With the Nondominated Sorting Genetic Algorithm With an Elite Strategy 2 and by Comparing Different Moving Strategies
Document Type
Periodical
Source
IEEE Intelligent Transportation Systems Magazine IEEE Intell. Transport. Syst. Mag. Intelligent Transportation Systems Magazine, IEEE. 15(3):82-100 Jun, 2023
Subject
Transportation
Aerospace
Computing and Processing
Components, Circuits, Devices and Systems
Fields, Waves and Electromagnetics
Layout
Space exploration
Optimization
Space vehicles
Transportation
Turning
Sensitivity
Language
ISSN
1939-1390
1941-1197
Abstract
Fully self-driving cars, or autonomous vehicles (AVs), will have self-parking functions, which will revolutionize and reshape parking facilities in several ways. Automated parking technology could realize high-density and intensive parking space efficiency through multirow layouts and with narrower spaces per vehicle, which would completely change traditional parking facilities. Nonetheless, extra movements and congestion are generated by releasing stuck vehicles inside rows of parking spaces. Profits in parking space efficiency might be reduced by losses extra movements and congestion. This article aims to optimize the multirow and multichannel layout of parking facilities for AVs, balancing maximal space utilization and minimal movements while satisfying a designated parking demand. A new moving strategy, “first storage and shortest moving to outdistance (FSSMD),” is put forward first. Then, the geometric attributes, other movement strategies, demand characteristics, and other design elements of parking facilities are comprehensively quantified to describe the profit and loss of self-parking. An optimization model for the multirow layout of parking spaces is constructed with the dual goals of maximizing the number of berths and minimizing the number of movements. The nondominated sorting genetic algorithm with an elite strategy 2 is designed to solve the optimal model. Finally, the optimal zones and layouts of parking facilities are put forward by numerical experiments, and the sensitivity of different moving strategies is discussed. The results show that the moving strategy based on the FSSMD significantly reduces expected movement times, without decreasing the utilization rate of parking spaces. Although having only one zone enables maximizing the number of parking berths and the utilization rate of parking facilities, such benefits and improvements might be compromised by additional movements. Therefore, the multirow optimal layout with a zoning of 4 and 2 interzone channels should be decided by assessing the paradoxical relationships between the maximal utilization rate and minimal movements.