Scholarly article on topic 'Parking Pricing for a Sustainable Transport System'

Parking Pricing for a Sustainable Transport System Academic research paper on "Economics and business"

CC BY-NC-ND
0
0
Share paper
Academic journal
Transportation Research Procedia
OECD Field of science
Keywords
{"parking pricing" / "limited traffic zone" / "parking zones" / "park and ride" / "shuttle bus."}

Abstract of research paper on Economics and business, author of scientific article — Marco Migliore, Antonino Lo Burgio, Manlio Di Giovanna

Abstract The purpose of this study has been the develop of a model for designing an efficient parking pricing policy. The aim is an intelligent control and management system of parking pricing integrated with a redefinition of the circulation scheme for a limited traffic zone in the Central Business District (CBD) of Palermo. The transport demand over the entire area of the town has been studied in order to design various parking pricing scenarios with the application of an additional cost on parking inside the selected area of the CBD. This area attracts most of the private vehicular traffic and it is characterized by university faculties, schools, hospitals, offices and commercial areas. The optimal hourly toll is defined by an iterative maximization process of an objective function. This objective function is subject to the following constraint: the percentage of available parking in the various parking zones has to remain major of the 30%. In this way, the users who need to park close to their final destination can easily find parking. Otherwise it is possible to leave the private car in a “park and ride” area and taking a shuttle bus directed towards the zones of the CBD. A basic principle of this pricing policy is the re-use of revenues for two purposes: to design a shuttle bus service that connects the various “park and ride” areas to the CBD and to improve the local public transport service on the OD pairs that show high travel demand. At the same time it is necessary to eliminate the stop and go flow in cordon roads to increase the capacity and avoid congestion in these critical links. The method shows that in a very simple, and relatively fast, way is possible to get a proposal for the modification of the parking pricing scheme that makes the city center no longer stifled by private car traffic.

Academic research paper on topic "Parking Pricing for a Sustainable Transport System"

Available online at www.sciencedirect.com

ScienceDirect

Transportation Research Procedía 3 (2014) 403 - 412

Transportation

Procedía

www.elsevier.com/locate/procedia

17th Meeting of the EURO Working Group on Transportation, EWGT2014, 2-4 July 2014,

Sevilla, Spain

Parking pricing for a sustainable transport system

Marco Migliorea*, Antonino Lo Burgioa, Manlio Di Giovannaa

aUniversity of Palermo - Transport Research Group, Viale delle Scienze edificio 8, Palermo 90128, Italy

Abstract

The purpose of this study has been the develop of a model for designing an efficient parking pricing policy. The aim is an intelligent control and management system of parking pricing integrated with a redefinition of the circulation scheme for a limited traffic zone in the Central Business District (CBD) of Palermo.

The transport demand over the entire area of the town has been studied in order to design various parking pricing scenarios with the application of an additional cost on parking inside the selected area of the CBD. This area attracts most of the private vehicular traffic and it is characterized by university faculties, schools, hospitals, offices and commercial areas. The optimal hourly toll is defined by an iterative maximization process of an objective function. This objective function is subject to the following constraint: the percentage of available parking in the various parking zones has to remain major of the 30%. In this way, the users who need to park close to their final destination can easily find parking. Otherwise it is possible to leave the private car in a "park and ride" area and taking a shuttle bus directed towards the zones of the CBD. A basic principle of this pricing policy is the re-use of revenues for two purposes: to design a shuttle bus service that connects the various "park and ride" areas to the CBD and to improve the local public transport service on the OD pairs that show high travel demand. At the same time it is necessary to eliminate the stop and go flow in cordon roads to increase the capacity and avoid congestion in these critical links.

The method shows that in a very simple, and relatively fast, way is possible to get a proposal for the modification of the parking pricing scheme that makes the city center no longer stifled by private car traffic.

© 2014TheAuthors. Published byElsevierB.V.Thisis an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/3.0/).

Selection and peer-review under responsibility of the Scientific Committee of EWGT2014 Keywords: parking pricing, limited traffic zone, parking zones, park and ride, shuttle bus.

* Corresponding author. Tel.: +39 091 23842416; fax: +39 091 423105.

E-mail address: marco.migliore@unipa.it, antonino.loburgio@unipa.it, manlio.di_giovanna@unipa.it

2352-1465 © 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/3.0/).

Selection and peer-review under responsibility of the Scientific Committee of EWGT2014 doi: 10.1016/j.trpro.2014.10.021

1. Introduction

For several years, the private car has been the privileged means by citizens to respond to the challenge of urban mobility. The private vehicular traffic invades daily the urban centers, contributing significantly to worsen the living conditions of the city. Therefore, it is important to supply innovative and sustainable alternatives.

Primarily, we should improve the supply of public transport in terms of level of service. At the same time, we should restrict access and use of private vehicles. Unlike other pricing policies, the parking pricing has very low investment costs but management costs very high, because the control and monitoring systems are mainly based on the use of staff.

We propose an innovative pricing model, from the point of view of technology and methodology, taking into account technical and economic feasibility of the mobility and of the environmental impact.

This study proposes a control system and intelligent management of the rates of parking spaces, such as alternative solution to the current pricing policy within the LTZ (Limited Traffic Zone). An important part of the proposal is the methodology for the calculation and optimization of the parking fees. The problem is solved with an heuristic approach that allows you to find the solution through the optimization of an objective function subject to constraints.

A basic principle of this pricing policy is the re-use of revenue for two main purposes: to strengthen local public transport systems on the OD pairs that show high travel demand and to incentivize the park and ride system.

This methodology has been applied to the city of Palermo to highlight the potential and limitations of the model. For the application of the method, it has been used free and open source software. In particular, it has been used the open source software for spatial analysis QGIS (Quantum GIS). It has allowed building the graph and analyzing the results. Finally, the free application AxE was used for the assignment of the transport demand to the supply.

2. Pricing model proposed

The pricing model proposed has all the features of a parking pricing scheme, but with the potentiality to affect the mobility inside the area of interest following the same goals and obtaining the same results of a traditional cordon pricing. This final goal can be obtained by means an appropriate traffic scheme inside the area of interest, with a smart control and management of the parking inside such an area. First of all, it is necessary to redesign the circulation scheme to cut out the crossing traffic form LTZ and to reduce the parking search time. The crossing traffic is the component of traffic related to that widespread behavior of leveraging the infrastructures of an area to pass through the area to reach a place located at the opposite corner of the entry's one. This component can be deviated on the perimeter of the safeguard area.

We apply the circulation rooms inside the area, that makes impossible to go through one to the opposite corner of the area, but it forces vehicles to go out from the area just by the adjacent room. In this way, a vehicle entering in the area has the only objective to reach a destination point inside the "room" from which it entered and to park there. The pricing phase will take place exactly at the moment of the parking, such pricing will be related to the fact of being inside the LTZ, indeed it could depend on a series of factors, like time slot (peak or soft), exact position of park and different covered distance from the gate, day of the week and others parameters that local government can take into account.

□□□□□DDDDDDD

□□□□□□ □□□□□□ □□□□□□ □□□□□□ □□□□□ □□□□□□ □□□□□□ □□□□□□ □□□□□□ □□□□□□

□□□□□□ □□□□□□ □□□□□□ □□□□□□ □□□□□ □□□□□ □□□□□□ □□□□□□ □□□□□□ □□□□□□ □□□□□□

Fig. 1. Scheme for creation of circulation rooms

In turn, within each room is proposed a circulation of rooms: unique ways of opposing both roads that converge at an intersection. In general, this type of movement breaks the continuity of the one-way roads that intersect only at intersections generating trajectories that flow in or flow out, so increasing the safety of the intersections. It will occur even less time to find a parking.

Fig. 2. Scheme of circulation in rooms

The critical question is now related to monitoring and manage the park system, that represent the tool able to reach two goals: the first is to create a LTZ but indirectly having also a more effective control about park.

Once identified the parking zones inside the area object of the intervention, will be installed a wireless sensor network (each park will be endowed with one sensor), which will allows monitoring in real time each park's state. Our Network is made by three kinds of nodes:

• Wireless sensor;

• Vehicular Tag;

• Gateway.

The wireless sensor's functionality is to recognize if park is busy or not and if the vehicle is equipped with an identifier, in this case the Vehicular Tag; it will periodically communicate with the Gateway to let system know its associated park's state. In the event when park is occupied by a vehicle without any Vehicular Tag, it will report the busy state to the Gateway, which in turn will inform central system about. Wireless sensor is equipped with a magnetometer and a wireless trans receiver, it monitor and record the perturbation of Earth's magnetic field, and send such records to the nearest Gateway, delegated to collect this data. The electronics board have to be embedded in an aluminum/plastic case which allows to reach a good mechanics strength and make sensor carriage able. This sensor will have shape and dimension of a well-known road marker used for horizontal signs, and will be installed on the ground by means of two-component glue.

Wireless sensor is managed by a CPU able to process the records about Earth's magnetic field, and thanks to wireless network interface to exchange these data with others sensors. Each sensor is powered by a battery charged by a little amorphous photovoltaic panel positioned on the marker's upper face, with an estimated life of 5 years.

The Vehicular Tag is a passive radiofrequency transponder without any battery, but endowed with just a chip and an antenna which task is double, receive and transmit data and transform electromagnetic power to supply voltage for the transmitting circuit. The smart side of each Tag is made just by a signal transmission circuit and by a nonvolatile memory which hold an univocal code to be transmitted to the reader when queried. The dimensions of the Tag are quite smalls and allow to realize it as a smart card, so it can be placed inside the vehicle to identify the owner, functionality useful to know the membership group and then his kind of pricing; for example people with disability, or unloading , resident, subscribers. Any other user without a Tag, once occupied the park, has a defined default amount of time (10 minutes for example) to pay for the parking, before it will be signaled as unauthorized busy; the payment will be possible following several choices: smartphone app, sending an sms with the number of the occupied stall and then paying by phone credit, prepaid cards equipped with Interactive voice response IVR, etc.

The Gateway or sink node is the network's node delegated to collect data from the sensors and to transmit this to a remote central server trough a technology among GSM/GPRS, MAN (Metropolitan Area Network), WAN (Wilde Area Network) or Power line leveraging the existing public power grid; remote server will then instantly update the global information.

Fig. 3. Disposition sensors and gateway in the parking area

Summarizing system's working is simple: thousands of parks equipped with wireless sensors on the ground, which collect instantly information about the availability of each related park, with relational authorization, and transmit this information to remote central server through nearest Gateway. All the collected data are managed in several ways:

• by the central server, which so has a detailed overview of the state of parks in the LTZ and can then modify pricing in each subzone by running custom made algorithms based on machine learning theory;

• by the verifiers/checkers, which by means of tablet have the exact position of the park illegally occupied and can then act rapidly and efficiently

• by users, which thanks to app for smartphone and tablet (Android iOS) will can find free parks and choose the shortest route to reach them

2.1 Optimization of the rate

The methodology for the calculation and optimization of the parking rate is a fundamental part of the proposal. Frequently, for the large number of variables involved in the optimization pricing problems, are used heuristic algorithms such as the algorithm proposed by D'Acierno et al. (2006) that helps to find great parking prices for each OD pair. Shoup (2004, 2006, 2007) proposes that the excellent rate can be found in an iterative way by an increase/decrease of the initial price. This principle was later applied in 2010 in San Francisco (SFpark) giving excellent results. Starting from the analysis of these works it was decided to formalize the problem of design of the parking fee as an optimization problem of an objective function subject to constraints.

Considering the two modes of transport (car and bus), the optimal price can be estimated by maximizing the

objective function:

Argmax ^ = Surplususers + Revenuepricing + Revenuebus — Costoperating + Externality (1)

Argmax(p) = j^-Zod dod [¿n(ew + ev»us)pricing - ln(ev™r + + (2a) +CP* 0Dcar_park(p) + ticketbus * ODnewpass_bus(P)

-4*Nbus(p)*l0+ (2c)

+0,11 * 0Dnewpass bus (p) * kmbus (2d)

subject to the constraints:

p>0 (3)

slotsfree > L (4)

/* =Al. (ccar (P- /*)) dcar (0far (P- /*) - 01bus (P)) (5)

where:

• the term "Surplususers" (2a) shows the user's "surplus", expressed as the product of the total demand dod and the variation in satisfaction between the project scenario and the initial one (with zero price), then divided by the coefficient pcost to express the surplus in Euros (€); the satisfaction has been calculated using the logsum function (user maximum perceived utility) for each origin-destination o-d pair (Vcar and Vbus are the car systematic utility function and the bus systematic utility function);

• the terms "Revenuepricing" and "Revenuebus" (2b) shows the revenues of the pricing due to two contributions: the pricing (CP) paid by cars (ODcar_park are the private vehicle users for each origin-destination) and the bus ticket (ticketbus) paid by the users who move from the private to the public transport due to the pricing (ODnewpass_bus);

• "Costoperating" (2c) shows the additional operating costs due to the expansion of public transport that is estimated as: 4*Arj„J*10 considering that the cost for each km/h of speed is equal to € 4 for a bus that moves at a medium speed of 10 km/h (Nbus, number of buses);

• the term "Externality" (2d) shows fewer externalities due to the transfer of a part of the demand from the private car to the bus (kmbus, distance traveled by bus). The difference of the two externalities coefficients are estimated as (Amoroso S. et al., 2012; European Environment Agency, 2005):

Finally, the three constraints respectively indicate that the pricing and the bus ticket (p) has to be positive (3), the percentage of free parking slots (slotsfree) has to be greater than or equal to a certain predetermined value (L, e.g. 1530%) (4) and the flow on each link of the network must be the solution of the fixed point problem depending from the interaction between the transport supply function and the transport demand function (5). In the last constraint f* is the equilibrium flow vector, g is the parking pricing vector, A is the incidence link-path matrix, £ is the path choice probability matrix, d^ is the car demand vector, C^ and Cbus are the costs by car and by bus for each path. The constraint on the percentage of free parking slots has a very important meaning because the verification of the same gives the user the certainty of being able to park within the area. It has a significant impact on the optimization of the model. As suggested by previous studies, the percentage of free stalls should not be less than 15% (Vickrey, 1954). For this problem, the entire process of determining the OD matrix, modal choice and assignment, can be solved with an algorithmic approach; however, the determination of the optimal rate of parking is resolved by an heuristic process, which will be explained below.

0,11 ipasLkm) * newPassbus (Ef£) * kmbus(km)

.pass.*km.

Modal split model

Start/Stop

Processing

Reading/Writing

Control

Fig. 4. Flowchart of the heuristic

In the various iterations the rate increase or decrease by € 0.25 or € 0.50, depending on the percentage of fill or saturation of the parking area. In the optimal scenario it will record the higher value of the objective function.

A multinomial logit model was used to simulate the choice among car, bus and train, with the systematic part of utility functions, a socioeconomic attribute and a psychometric indicator (Migliore et al., 2013). The variables involved in the model are: walking time (min.), waiting time (min.), in-vehicle time (min.), parking time (min.), cost (€), high-grade worker (entrepreneur, manager or a professional) and discomfort.

Table 1. Model specification

^^Ceff. Alt^ ASCtus ASCc/_ ar ßT-walk ßT-wait ß'T-in vehicle ^ T-parking ft Cost ^'HG-worker ^Discomfort

Car 0 0 twalk 0 tin vehicle tparking cost 0 0

(min.) (min.) (min.) (€)

Bus 1 0 twalk twait tbus 0 cost 0/1 1^5

(min.) (min.) (min.) (€)

ASCbus (Car): alternative specific constant for the bus (car) option; HG-Worker:\ if the respondent are an high-grade worker, 0 otherwise; Discomfort/Dirt: perceived importance of the discomfort and dirtiness characterizing public vehicles (crowded, not always air-conditioned, dirty) discomfort attribute. T-Walk\ walking time attribute; T-Wait: waiting time attribute; T-In Vehicle: in-vehicle time attribute; T-Parking: parking time attribute; Cost: transport cost variable.

The mode choice behaviour was simulated referring to random utility theory; in particular, the empirical analysis was performed through the logit framework (Ben Akiva M. et al, 1985). To enhance the explanatory power of the mode choice model, a base specification was integrated with a socioeconomic characteristic of the respondents and with an indicator of their psychological traits deriving from the responses to survey questions on individual attitudes and perceptions (Migliore M. et al., 2012). One thousand and two hundred SP choice observations were processed for model estimations, which were executed by the BIOGEME 1.8 econometric software (Bierlaire M., 2009).

Table 2. Estimated models

Coefficients of modes' attributes Value t-test p-value

ASCBub -0.551 -1.45 0.15

ft T-Walk -0.0207 -2.40 0.02

B r T-Wait -0.0132 -0.75 0.45

ft'T-in vehicle -0.0196 -3.40 0.00

^T-parking -0.0362 -1.97 0.05

ft Cost -0.218 -2.73 0.01

ft HG-worker -1.36 -6.26 0.00

ft comfort -0.153 -2.44 0.01

Final log likelihood =-459.259

P = 0.582

Adjusted P = 0.574

3. The parking pricing optimization in the historical centre of Palermo

An application on the CBD of Palermo has been studied using free software or open source: the OpenOffice Calc spreadsheet for data entry and data processing, the software AxE for the assignment (Assignment for Experts) made by Prof. Giorgio Salerno and Quantum GIS (QGIS) for the modeling of the road graph and for the graphic elaboration of the results obtained by the assignments.

Nowadays the parking system is regulated by the blue areas, with variable tickets with an average of 1 €/hour. There are no restrictions on access to private vehicles, and the local public transport system has a very low attractiveness due to the low level of service (due to the low values of service frequency and low values of running

speed - about 8 km/h without reserved lane). In fact the modal split is strongly in favor of a private vehicle with a percentage of 82% compared to 18% of public transport. This is the real scenario which flows on the network are shown in the figure below.

Fig. 5. Saturation on the network in real scenario

From this initial situation start the heuristic process described by the flowchart above. The change, at each iteration, of the parking rate involves the restatement of the parameters through the assignment and consequently the estimation of the modal split. The various iterations have been carried out by varying (increasing or decreasing) the parking ticket by € 0.25, € 0.50 or € 1.00 for each parking area, depending on the load.

The heuristic has been initiated with the introduction of the first rate variation, follows the assignment of demand to supply through the software AxE.

At the end of the iterative process, the estimation of the objective function and its relative constraint have been calculated. Starting from the real scenario with a constant rate (1 €/h) for all the parking slots we simulated the increase and the decrease of the ticket for every parking area depending on the percentage of fill or saturation. This iterative process continues until the objective function takes the maximum positive value and the constraints are respected.

In this application, we have performed two sets of iterations. In the first set there are the variations of the tickets and the design of a shuttle bus service that connects the various "park and ride" areas to the CBD, in order to finance an alternative transport mode for the users that do not want to pay the parking pricing in the CBD. We performed 12 iterations, the values of the objective function are shown in the diagram below; when the constraints are verified the iterations are highlighted in green.

Fig. 6. Objective function in the first set

The eleventh iteration presents the maximum value of the objective function and the constraints are verified. The occupancy rate of the slots of the entire area is about 0.70 (a percentage of free slots of 30%). However, the percentage of bus users is 0.21, and it grows to 0.25 if we introduce the park and ride solutions.

In the second set of iterations we introduced a reorganization of the intersections and eliminated the possibility of stopping in cordon roads to increase the capacity and avoid congestion in these critical links. Then the revenues has been used for improving the local public transport service on the OD pairs that shows high travel demand.

We performed 6 iterations. As shown in the diagram below, initially the objective function has a very high positive value, due to the two variations introduced: the reinforcement of the public transport service and the improvement of the cordon roads; but the constraints are not verified. After the increase in parking fees, there was a recovery, reaching a relative maximum at the fifth iteration. The optimal configuration is the fifth iteration in which the bus users is about the 21 percent, and they grows to 25 percent if we introduce the park and ride solutions.

40000 30000 20000 10000 0

-•— F.O.

""m -•- --•

0 1 2 3 4 5 6

30145 18444 9557 12810 13760 11756

Fig. 7. Objective function in the second set

It is important to underline that the improvement of the cordon roads has allowed a reduction and a better rationalization of fees within the park, with better detection of parking pricing zones for the benefit of the citizens.

Fig. 8. Parking pricing zones (tickets in euros)

4. Conclusions

The two sets of simulations have shown that a strategic choice of design can have a significant impact on the modal split and on the objective function, leading to two solutions quite different from each other.

Using the pricing to increase the sustainability of the system, means use the cash flow arising from pricing to finance the transportation alternatives that would lead user to use the Public Transport and the "park and ride".

In addition, if the method is used in the operational phase, with the data collected in the monitored areas, it allow to calibrate the optimal rate more precisely.

Introduce appropriate parking fees allows you to decongest the areas that you want to leave by private vehicles. However, this requires that the fees are reformulated periodically, as in the case of SFpark, to avoid the displacement of the phenomenon of congestion in other areas.

The results of the application shows that in a very simple, and relatively fast way, is possible to get a proposal for the modification of the parking pricing scheme. These results could make the city center no longer stifled by private car traffic.

References

Amoroso, S., Catalano, M., Galatioto, F., Migliore, M. (2012) The interaction between rail stations and urban area in medium-sized towns. A

case study for Trapani, Ingegneria Ferroviaria, LXVII, 11, 923-941. Amoroso, S., Migliore, M., Catalano, M., Galatioto, F. (2010) A demand-based methodology for planning the bus network of a small or medium

town, European Transport, Trasporti Europei, 44, 41-56. Ben Akiva, M. E. & Lerman, S. R. (1985) Discrete Choice Analysis: Theory and Application to Travel Demand, The MIT Press: Cambridge, Mass.

Bierlaire, M. (2009) Estimation of Discrete Choice Models with BIOGEME 1.8 , Transport and Mobility Laboratory, EPFL École polytechnique fédérale de Lausanne.

Canterella, G.E., Introduzione alla tecnica dei trasporti e del traffico con elementi di economia dei trasporti, UTET, 2001. Cascetta E., Teoria e metodi dell'ingegneria dei trasporti, UTET 1998.

Catalano, M., Lo Casto, B., Migliore, M. (2008) Car sharing demand estimation and urban transport demand modelling using stated preference

techniques, European Transport, 40, 33-50. D'Acierno L., Gallo M., Montella B. (2006), Optimisation models for the urban parking pricing problem, Transport Policy 13, pp.34-38. European Environment Agency (2005) Indicator fact sheet - TERM 2005 25 :EEA, Copenhagen, Denmark.

Migliore M. (1998), Un modello d'ottimizzazione domanda-offerta nella sosta, Proc. 3rd Conf. Rilievi, modellizzazione e controllo del traffico veicolare, Rende (Italy), 1998.

Migliore M. (2002), The perception of network congestion by using parking fees, Urban Transport VIII. ISBN: 1-85312-905-4

Migliore M., Catalano M., Castelluccio F., Lo Burgio A., Di Giovanna M. (2013), The latent factors behind the urban travel behaviour,

EWGT2013, 16th Meeting of the EURO Working Group on Transportation, Porto, Portugal, 4-6 September 2013. Migliore, M., Catalano, M., Lo Burgio, A., Maritano, L. (2012) The analysis of urban travellers' latent preferences to explain their mode choice

behaviour, WIT Transactions on Ecology and the Environment, 162, 193-203. Pigou A.C. (1920), The Economics of Welfare, MacMillan London. Salerno G., manuale programma di assegnazione AxE (Assegnazione per Esperti).

Shoup D. (2004), Curb Parking: the ideal source of local public revenue, Regional Science and Urban Economics, 34, 6, 753-784. Shoup D. (2006), Cruising for parking, Transport Policy, 13(6), 479-486. Shoup D. (2007), Cruising for Parking, Access, No. 30, Spring 2007, pp.16-22.

Vickrey W. (1954), "The economizing of curb parking space", Traffic Engineering, November 1954, 62-67.

Vickrey W. (1992), Principles of Efficient Congestion Pricing, Columbia University.

Yosef Sheffi (1985), Urban Transportation Networks, Prentice-Hall, Englewood, 1985, pp. xvi-399.