智能停车场论文翻译英文.doc

Intelligent parking system

Abstract
The basic concepts of the parking reservation system and parking revenue management system are discussed in this paper. The proposed intelligent’’ parking space inventory control system that is based on a combination of fuzzy logic and integer programming techniques makes ‘‘on line’’ decisions whether to accept or reject a new driver's request for parking. In the first step of the proposed model, the best parking strategies are developed for many different patterns of vehicle arrivals. These parking strategies are developed using integer programming approach. In the second step, learn-ing from the best strategies, specific rules are defined. The uniqueness of the proposed approach is that the rules are derived from the set of chosen examples assuming that the future traffic arrival patterns are known. The results were found to be close to the best solution assuming that the future arrival pattern is known.

Keywords: Traffic; Uncertainty modeling; Control; Parking; Fuzzy logic
1. Introduction
Every day a significant percentage of drivers in single-occupancy vehicles search for a parking space. Additionally, less experienced drivers or out-of-towners further contribute to the increase of traffic congestion. Search for a vacant parking space is a typical example of a search process. Every parking search strategy is composed of a set of vague rules. It is usually difficult to describe these rules explicitly. The type of the planned activity, time of a day, day of the week, current congestion on particular routes, knowledge of city streets, and potentially available parking places have signficant influence on a chosen parking search strategy. During the last four decades numerous parking search models have been developed (Vander Goot, 1982; Axhausen and Polak, 1991; Polak and Axhausen, 1990; Young et al., 1991a,b; saltzman, 1997; Shoup, 1997; Steiner, 1998; Thompson and Richardson, 1998; Arnott and Rowse, 1999; Tam and Lam, 2000; Wong et al., 2000; Waterson et al., 2001). In many decision-making situations in transportation (modal split, choice of air carrier, choice of airport, etc.) the competitive alternatives and their characteristics are reasonably well known in advance to the decision maker (passenger, driver). On the other hand, the drivers usually discover diffierent parking alternatives one by one in a temporal sequence. Clearly, this temporal sequence has a very strong influence on the driver's final decision about the parking place.
During the past two decades, traffic authorities in many cities (Helsinki, Cologne, Mainz, Stuttgart, Wiesbaden, Aalborg, Hague) have started to inform and guide drivers to parking facilities with real-time var-iable message signs [directional arrows, names of the parking facilities, status (full, not full, number of available parking spaces, etc.)]. Information about the number of available parking spaces could be displayed on the major roads, streets and intersections, or it could be distributed through the Internet.
It is logical to ask the question about the benefits of the parking
guidance systems. Current practice shows that parking guidance systems usually do not change the occupancy rate or average parking duration. Drivers easily become familiar with the parking guidance systems, and majority of them use, thrust and appreciate the help of the systems. Guidance systems significantly increase the probability of finding vacant parking space, mitigate frustration of the drivers–visitors unfamiliar with the city center, decrease the queues in front of parking garages, decrease the total amount of vehicle-miles traveled (particularly in the city centers), decrease the average trip time, energy consumption, and air pollution. Parking guidance system is a part of comprehensive parking policy and traffic management system, whose other elements are street parking control (including sanctions for the illegally parked vehicles), parking fare structure, and parking revenue management system.
Parking guidance systems help drivers to find vacant parking spaces when they are already on the network, and approaching their final destination. Throughout this research the concepts of the parking reservation system and parking revenue management system are proposed. Such systems would help drivers to find a vacant parking space even before beginning their trip. The proposed “intelligent” parking space inventory control system that is based on the combination of simulation, optimization techniques, and fuzzy logic makes “real-time” decisions as to whether to reject or accept a new request for parking. The proposed methodology could be applied for parking lots and parking garages in cities and at the big international airports.
The paper is organized as follows: Parking-pricing problems are presented in Section 2. Analogies between parking problems and some other industries are presented in Section 3. The parking revenue management system is introduced in Section 4, and the Intelligent parking space inventory control system is introduced in Section 5. The algorithm to create intelligent parking spaces inventory control system is presented in Section 6. Results obtained with the “intelligent” parking system are given in Section 7, and Sec-tion 8 presents the concluding remarks and further research orientations.
2. parking pricing
In majority of cities throughout the world drivers pay for using different parking facilities. In some instances, traffic congestion can be significantly reduced as a result of parking price. The parking revenue is usually used to cover parking facility costs (access gates, ticket printers, parking meters, parking signs, attendants), or to improve some other traffic and transportation activities. Different parking pricing strategies should be a part of the comprehensive solution approach to the complex traffic congestion problems. There is no doubt that parking pricing represents one of the important demand management strategies. For
example, traffic authorities, local governments and private sector could introduce higher parking tariffs for solo drivers or for long-term parkers in congested city areas. They could provide special parking discounts to vanpoolers. Obviously parking pricing should be carefully studied in the context of the considered city area (down-town, residential, commercial, retail use areas).
In some cities (Madison, Wisconsin) there are already time dependent parking fees that force commuters to switch to different alternatives of public transportation. Trying to promote public transit San Francisco traffic authorities increased parking tariffs at public and commercial garages. The Chicago authorities raised parking rates few times. As a consequence, the total number of cars parked significantly decreased, as well as parking duration time. The greatest decrease was in the number of all day parkers. Authorities in Seattle signficantly reduced parking tariffs for carpool at two Seattle parking facilities in downtown. Active role in parking pricing strategies could also have employers paying for employees' parking. Employers who remove parking subsidies for the employees could significantly decrease the total number of solo drivers. The main role of any parking pricing strategy should be reducing the total number of vehicle trips during certain time periods, shifting commuters to alternative transportation modes, and to different parking locations. At the same time, when trying to implement any parking strategy, it is very important to provide enough parking space for shoppers, to provide preferential parking for residents in considered city area, to provide preferential parking for different parking locations, to consider low income families, and to protect streets in the neighborhood from illegal parking.
The basic economic concepts of supply and demand should be more utilized when solving complex traffic congestion and parking problems (Vickrey, 1969, 1994; Verhoef et al., 1995). So-called value pricing is also known as congestion pricing, or variable tolling. The basic idea behind the concept of congestion pricing is to force drivers to travel and use transportation facilities more during off-peak hours and less during peak hours.
The idea of congestion pricing is primarily connected with the road (drivers pay for using private, faster roads, drivers with lower vehicle occupancy pay for using High Occupancy Vehicle lanes, drivers pay more to enter city's downtown on weekdays) or airport operators (more expensive landing fees during peak hours). In the context of parking problems, this means: (a) that different parking tariffs should exist for different users; (b) that the parking fees should increase and/or decrease few times during a day.
3. Parking problems and revenue management systems: Analogies with some other industries
Airline industry, hotels, car rental, rail, cruise, healthcare, broadcast industry, energy industry, golf,equipment rental, restaurant, and other industries are utilizing revenue management concepts when selling their products (Cross, 1997). Revenue management could be described as a group of different scientific techniques of managing the company revenue when trying to deliver the right product to the right client at the right price at the right time. The roots of the revenue management are in the airline industry. The basic characteristics of the industries to which different revenue management concepts were successfully applied are: (a) variable demand over time; (b) variable asset utilization; (c) perishable assets; (d) limited resources; (e) market segmentation; (f) adding new capacity is expensive, difficult or impossible; (g) direct cost per client is negligible part of the total cost of making service available; (h) selling products in advance. The main characteristics of the parking space inventory control problems are the following:
· Parking demand is variable over time.
· Like hotel rooms, or restaurant chairs, parking spaces also have daily opportunity to be ‘‘sold’’ (used by clients).
· Any parking lot or garage has limited number of parking spaces that can be used by drivers
· Market segmentation means that different customers are willing to pay different prices for the same asset (hotel room, airline seat, seat in a rented car). Businessman wanting to park a car near a meeting point 15 minutes before the meeting would be ready to pay much higher parking fee than a pensioner planning to walk with his wife through the downtown, who made parking reservation four day in advance.
· Building new garages and parking lots could be very expensive and sometimes very difficult.
· Parking places can be easily reserved in advance.
Introducing and developing parking reservation system (created in an Internet and cell phone environ-ment) would present further improvement in modern parking technologies. Drivers would be advised and guided before beginning of the trip, as well as during the trip. Parking reservation system should be coupled with the parking revenue management system. In this way, parking operators and traffic authorities would be able to implement different parking strategies. Once the driver is allowed to park, it is possible to implement internal garage guidance system that guides the driver to an empty parking place.
4. Introducing parking revenue management system
Let us assume that we have parking reservation system. Drivers make their requests for parking at random moments of time (by phone from home, by cell phone while driving, through the Internet, etc.).
A certain number of drivers would maybe cancel their reservations before beginning of the parking.These cancellations would also be made at random moments of time. Like in some other industries, a certain number of drivers would not appear in parking garage for which they have a con-firmed reservation and purchased ticket. Would these drivers be penalized for their behavior? Depending on ration between parking demand and parking supply, the answer could be ‘‘Yes’’ or ‘‘No’’.
Reservation system should be flexible enough allowing some drivers to appear right before wished beginning of parking, looking for an empty space in a garage, even though they do not have a confirmed reservation. Would it be good to have few different parking tariffs? The answer is obviously ‘‘Yes’’. Drivers paying lower parking tariffs could be disabled and senior citizens, people who reserve parking space few days in advance, or HOV drivers. Drivers paying higher tariffs could be solo drivers, long term parking drivers, or drivers showing up and asking for parking without making reservation in advance. Obviously, there is a lot of possible parking pricing strategies.
The stochastic nature of reservation generation and cancellation, the stochastic nature of driver show-up during reserved time slot, variety of parking tariffs, and the need to respond to drivers' requests in real time, indicate that the management of parking garage revenues represents a complex problem.
In the past 30 years a relatively large number of papers have been devoted to different aspects of the air-line seat inventory control problem (Littlewood, 1972; Belobaba, 1987; Brumelle and McGill, 1993; Teodorovic et al., 2002). The model proposed in this paper is highly inspired by the developed airline yield management stochastic and/or deterministic models.
Let us assume that we have few different parking tariffs. The simplest reservation system (similar to some airline reservation systems in the past) could be ‘‘distinct tariff class parking space inventories’’ (Fig. 1(a)),indicating separate parking spaces in the garage for each tariff class. In this case, once the parking space is assigned to a tariff class, it may be booked only in that tariff class or else remains unsold. There are certain advantages, as well as certain disadvantages in the case of distinct parking space inventories. In this case users paying lower tariffs would be relatively well ‘‘protected’’. In other words, this system would pay a lot of attention to the disabled person, senior citizens, people who reserve parking space few days in advance, and HOV drivers. Obvious disadvantage of the distinct parking space inventories is the fact that very often some parking spaces assigned to lower tariff users would be empty even the higher tariff users demand is very high. In other words, it would be possible to reject some drivers even all parking spaces in garage are not occupied.
Fig.1
(b)
(a)
BLm

BL2

BL1=C

C
BLm
BL2
BL1

In case of a ‘‘nested reservation system’’, the high tariff request will not be rejected as long as any parking spaces are available in lower tariff classes. For example, if we have four tariff classes, then there is no booking limit for class 1, but there a