Вы находитесь на странице: 1из 11

Proceedings of the 6th IMT-GT Conference on Mathematics, Statistics and its Applications (ICMSA2010) Universiti Tunku Abdul Rahman,

Kuala Lumpur, Malaysia

462

Linear Programming for Parking Slot Optimization: A Case Study at Jl. T. Panglima Polem Banda Aceh
Said Munzir1 , Mahyus Ikhsan1 and Zainal Amin1
1

Mathematics Department, Faculty of Sciences, Syiah Kuala University, Banda Aceh, Indonesia
smunzir2001@yahoo.com home page: http://www.math-usk.org/smunzir/

Abstract. This research investigate the optimization of the available parking area, based on parking requirement analysis and the identification of existing parking problem. The study is expected to provide information to reduce traffic delay due to on-street parking. It is related to the need of parking area while offstreet parking area is not available. For a case study, the street of Panglima Polem in Banda Aceh is chosen as it is located in the city central of Banda Aceh. The model use linear programming supported by observation and survey data to formulate optimization problem. The study has formulated a well-posed problem for parking slot optimization based on the user needs, interpreted as parking slot proportional to parking accumulation and duration. Keywords: parking slot optimization, linear programming

1. Introduction
Transportation plays an important and strategic role in the development of a nation, particularly in distributing the product of the development for all citizens. General problems that occur in urban transportation is the traffic jam. One of the sources of the traffic jam is the decreases of road diameter due to the use of the part of the road for on-street parking. This traffic jam has a massive effect if it is considered comprehensively. One of this effects is, for instance, the excessive use of fuels which cause a large amount of economic losses. Hence, the effort to reduce the traffic jam is necessary, one of which is the management of on-street parking problem. The traffic flow problem due to on-street parking is an extremely serious problem. Several researches has been conducted in relation to this particular problem, such as works of Sinaga [10] and Setiawan [8]. The analysis of parking demand based on characteristic of parking sites has been conducted by Yosritzal [11] and Widanengsih [12]. In addition, they also suggested that regression method can also be used to determine the standard of parking slot demand around a hospital.

Proceedings of the 6th IMT-GT Conference on Mathematics, Statistics and its Applications (ICMSA2010) Universiti Tunku Abdul Rahman, Kuala Lumpur, Malaysia

463

The parking management problems were also studied using transportation management science as performed by Rapp [6]. Cannon [2] develop a simulation program to handle the parking problem. On the other hand, model using research operation employing linear programming is also utilized by researchers such as Bruglieri [1], Cordone [3] dan Silva [9]. For long term desain Djakfar [4] suggested a method using criteria analysis. From a series of previous researches related to the parking problem management, we cannot find a research related to parking slot allocation using linear programming. The demand of parking slot is impossible to be fullfiled entirely especially if the strategy is to use the on-street parking, as the space is usually allocated for traffic flow. In reality, the problem that usually emerge in relation to on-street parking is to determine the best parking slot allocation to distribute among different types of vehicle on limited parking space. In this study, we focus on building a mathematical model describing the problem and try to allocate optimal parking slot proportional to each type of vehicle using linear programming. Peunayong area located in Banda Aceh is a center for bussiness activity in the city. The problem occurred in the road around the area is a heavily massive traffic activity, such as in the road segment of Jalan T. Panglima Polem. From simple observation, obvious sources of low road performance in this area is the on-street parking along the road and intersections. This on-street parking decreases road capacity and increase road side obstacle. This problem is the result of insufficient parking space available in the area.

2. Literature Review
Parking is defined as terminating a vehicle at a certain location and is a part of traffic circulation. Based on its location, the parking is classified into two categories, i.e.: on street parking and off street parking (Widanengsih dan Elkhasnet [12]). Parking Control Unit (PCU) is the parking space used for a vehicle, which depend on vehicle dimension plus additional space needed for a vehicle to maneuvre whose value depending on the parking angle. PCU of each vehicle can be obtained in the Table 2.1.

Proceedings of the 6th IMT-GT Conference on Mathematics, Statistics and its Applications (ICMSA2010) Universiti Tunku Abdul Rahman, Kuala Lumpur, Malaysia

464

Table 2.1 PCU of each type of vehicle Width (meter) 1 0,8 1,5 2,1 3,5 2,4 1,6 Parking Width (meter) 1,5 1,3 2,5 3,1 4,5 3,4 2,6 Length (meter) 2,2 1,9 4,1 6,0 9,3 7,2 4,1 Parking Length (meter) 2,7 2,4 5,1 7,0 10,3 8,2 5,1

No 1 2 3 4 5 6 7

Jenis Kendaraan Becak Motorcycle Passenger Car Medium Bus Big Bus Truck Small Bus

Parking characteristics are parameters related to the amount of parking demand that have to be provided. According to Hobbs [5], parking characteristics includes: a. Parking volume, i.e. number of vehicle entering a parking site. b. Parking accumulation, number of vehicle parked at a parking site at a certain time. c. Parking index, i.e. percentage of the vehicle occupied the parking area. d. Parking duration, i.e. time interval (in minute or hour) for a certain vehicle parked at a parking site. Percentage amount of parking duration is formulated as ratio between the amount of vehicle parked during certain time interval and total number of vehicle observed. e. Average parking duration, i.e. total number of vehicle parked during certain time interval compared to vehicle enter parking site. f. Parking exchanges, i.e. measurement of parking occupation calculated as ratio between the number of vehicle parked compared to parking capacity available. g. Parking utilization level, computed from the ratio between average parking and parking space capacity. Meanwhile, average parking is obtained from the ratio between sum of parking accumulation for all observation time and number of observation.

3. Research Method
Location of this survey is at Jln. T. Panglima Polem Peunayong Banda Aceh and was conducted whole day beginning from 07.00 a.m. to 18.00 p.m.. Secondary supporting data and information for this research is obtained from Dinas Perhubungan Provinsi Aceh. The method used to obtained parking space geometric data and existing parking space capacity is through measuring the parking space area and parking space allocation for each vehicle. Data sampling for vehicle entering or leaving the parking site along with its parking

Proceedings of the 6th IMT-GT Conference on Mathematics, Statistics and its Applications (ICMSA2010) Universiti Tunku Abdul Rahman, Kuala Lumpur, Malaysia

465

duration is obtained through a field survey walking along the parking site and counting the number of vehicle and its parking duration. Data processing is conducted using Microsoft Excel 2003 and QM for Windows 2.0. In relation to data obtained from the field survey, model analysis to be developed is conducted by including the user parking space demand proporsional to average parking accumulation and average parking duration. Proportionality of average parking accumulation is computed every 15 minutes for each type of vehicle. This proportion for motor cycle, car and becak are calculated using the following formulation:
p 1 x +x +x , 2 3 p +p +p 1 1 2 3 p 2 x +x +x , 2 3 p +p +p 1 1 2 3 p 3 x +x +x 1 2 3 p +p +p 1 2 3

where : Average parking accumulation of motor cycle (number of vehicle/15 minutes) p 2 : Average parking accumulation of car (number of vehicle/15 minutes) p 3 : Average parking accumulation of motor becak (number of vehicle/15 minutes)
p1

and x1 + x 2 + x3 : Parking space capacity

Mean while, proportionality to average parking duration every 15 minutes for motor cycle, car and becak are respectively formulated as: t3 t1 t2 (x1 + x2 + x3 ) (x1 + x2 + x3 ) (x1 + x2 + x3 ) t1 + t 2 + t 3 t1 + t 2 + t 3 t1 + t 2 + t 3 , , where
t1 t2 t3

: Average parking duration for motor cycle (minute) : Average parking duration for car (minute) : Average parking duration for becak (minute)

Proceedings of the 6th IMT-GT Conference on Mathematics, Statistics and its Applications (ICMSA2010) Universiti Tunku Abdul Rahman, Kuala Lumpur, Malaysia

466

The problem to solve is how to maximize parking space capacity at Jalan T. Panglima Polem subject to available parking land, and the same time meet the the demand of parking for each type of vehicle. The parking demand is based on proportionality of average parking accumulation and average parking duration. The focus here is to allocate parking space for all three types of vehicle. Based on parking space standard allocated for each type of vehicles by Dinas Perhubungan Aceh, parking space for each vehicle in this study is sebagai berikut: a. Parking space for motor cycle is 3,12 m2. b. Parking space for car is 12,75 m2. c. Parking space for becak is 4,05 m2. The structure of decision making for maximization of parking capacity can be arranged as:
Tabel 3.1 The structure of decision making for maximization of parking capacity

activity No Coefficient of objective functions 1 2 3 4 5 6 7 Parking space area Motor cycle parking accumulation proporsional to all Car parking accumulation proporsional to all Becak parking accumulation proporsional to all Proporsional waktu parkir rata-rata mobil Proporsional waktu parkir rata-rata becak motor
x1 c1 x2 c2 x3 c3

Limitation factor

a11 a31 a51 a61 a71

a12 a32 a52 a62 a72

a13 a 23 a33 a 43 a53 a63 a73

a21 a22 a41 a42

b1 b2 b3 b4 b5 b6 b7

Proportion of vehicle average parking accumulation every 15 minutes for each type of a vehicle is computed from the vehicle average parking accumulation divided by total average parking accumulation for all vehicles and then multiplied by parking space capacity. In mathematical model, it can be written as: 1. Proportion of average parking accumulation for motor cycle is
p1 (x1 + x2 + x3 ) p1 + p 2 + p 3

Proceedings of the 6th IMT-GT Conference on Mathematics, Statistics and its Applications (ICMSA2010) Universiti Tunku Abdul Rahman, Kuala Lumpur, Malaysia

467

2. Proportion

of

average

parking

accumulation

for

car

is

p2 (x1 + x2 + x3 ) p1 + p 2 + p 3

3. Proportion of average p3 (x1 + x2 + x3 ) p1 + p 2 + p3

parking

accumulation

for

becak

is

Where p1 is motor cycle average parking accumulation (number of vehicles in 15 minutes), p 2 is car average parking accumulation (number of vehicles in 15 minutes), p 3 is becak average parking accumulation (number of vehicles in 15 minutes), while ( x1 + x 2 + x3 ) is parking slot capacity to be allocated. Proportion of vehicle average parking duration every 15 minutes for each type of a vehicle is computed from the vehicle average parking duration divided by total average parking duration for all vehicles and then multiplied by parking space capacity. In mathematical model, it can be written as: 1. Proportion of average parking duration for motor cycle is
t1 (x1 + x2 + x3 ) t1 + t 2 + t 3 t2 (x1 + x2 + x3 ) t1 + t 2 + t 3 t3 (x1 + x2 + x3 ) 3. Proportion of average parking duration for becak is t1 + t 2 + t 3

2. Proportion of average parking duration for car is

Where t1 is average parking duration for motor cycle (minutes), t 2 is average parking duration for car (minutes), and t 3 is average parking duration for becak (minutes). 3.1. Processing Model in Linear Programming The problem of parking slot allocation at Jln T. Panglima Polem uses three parts of linear programming model, i.e.: Objective function:
Z = x1 + x 2 + x 3

Subject to constraints: 3,12 x1 + 12,75 x 2 + 4,05 x3 Parking space area

Proceedings of the 6th IMT-GT Conference on Mathematics, Statistics and its Applications (ICMSA2010) Universiti Tunku Abdul Rahman, Kuala Lumpur, Malaysia

468

x1

p1 (x1 + x2 + x3 ) p1 + p 2 + p3

p2 (x1 + x 2 + x3 ) p1 + p 2 + p3 p3 (x1 + x 2 + x3 ) x3 p1 + p 2 + p3 t1 x1 (x1 + x2 + x3 ) t1 + t 2 + t 3 t2 (x1 + x2 + x3 ) x2 t1 + t 2 + t 3 t3 (x1 + x2 + x3 ) x3 t1 + t 2 + t 3 And non-negativity constraints: x1 , x 2 , x3 , p1, p 2 , p3 , t1 , t 2 , t 3 0 0 1 and
x2

Table 4.1 Parking characteristics of Motorcycle, Car and Becak Parking Characteristics Parking volume Parking capacity Peak of parking accumulation: - Time Motor cycle 1211 147 14.30 14.45 p.m Car 421 16 11.30 11.45 a.m. Becak 100 9 08.45 09.00 and 09.0009.15 a.m 10 111.11 15 79.80 24,24 1.23 51.67

- Jumlah Kendaraan (kendaraan) Parking index (%) Parking duration: - Parking duration of largest number of vehicles (minutes) - Percentage of the number of vehicle park (%) Average parking duration (minutes) Parking exchanges Parking utilization (occupation) (%)

135 91.84 15 68.09 46,72 8.24 67.18

26 162,5 15 68.60 25,69 26.3 106.875

After substitution of the values of p1 , p 2 , p 3 , t1 , t 2 and t 3 from field survey into the model, the optimization problem can be written as:

Proceedings of the 6th IMT-GT Conference on Mathematics, Statistics and its Applications (ICMSA2010) Universiti Tunku Abdul Rahman, Kuala Lumpur, Malaysia

469

Maximize Z = x1 + x 2 + x3 Subject to constraints: 3,12 x1 + 12,75 x 2 + 4,05 x3 x3 Where x1 , x 2 , x3 , 0 and 0 1 . 588 0,82( x1 + x 2 + x3 ) 0,14( x1 + x 2 + x3 ) 0,04( x1 + x 2 + x3 ) 0,48( x1 + x 2 + x3 ) 0,27( x1 + x 2 + x3 ) 0,25( x1 + x 2 + x3 )

x1 x2
x3
x

x2

4. Results and Discussions


The model was initially tested for values of between 0.75 to 1 using QM for Windows 2.0 to find optimal solution. However, no feasible solution was found. This means that in trying to satisfy the average parking accumulation and parking duration simultaneously for each vehicle at more than 75% satisfaction, constraints cannot provide any feasible point in its interior. Following up these results, more realistic scenarios for the optimization were made, i.e.: 1. Optimization considering constraints only of average parking accumulation with values of from 0.75 up to 1. 2. Optimization considering constraints only of average parking duration with values of from 0.75 up to 1. 3. Optimization considering both constraints of average parking duration and average parking accumulation with values of from 0.5 up to 0.7.

4.1.

Formulation Considering Parking Accumulation Only

Optimization considering constraints only of average parking accumulation with values of from 0.75 up to 1 was run using QM for Windows 2.0 with the solution is shown in the Table 4.2.

Proceedings of the 6th IMT-GT Conference on Mathematics, Statistics and its Applications (ICMSA2010) Universiti Tunku Abdul Rahman, Kuala Lumpur, Malaysia

470

Table 4.2 Solution for Optimization considering average parking accumulation only Variable

x1 x2 x3
Z

= 0,75 121,6684
14,769 4,2197 140,6571

= 0,80 119,0373
15,575 4,45 139,0623

= 0,85 115,8873
16,2817 4,6519 136,8209

= 0,90 112,8731
16,966 4,8474 134,6505

= 0,95 109,8823
17,6289 5,0368 132,548

=1 107,0182
18,2714 5,2204 130,5101

From the solution in the Table 4.2, it shows that the higher the value of (level of satisfaction) the smaller the parking slot obtained especially the vehicle (motor cycle) having the highest accumulation in comparison to the others.

4.2.

Formulation Considering Parking Duration Only

Optimization considering constraints only of average parking duration with values of from 0.75 up to 1 was run using QM for Windows 2.0 with the solution is shown in the Table 4.3.
Table 4.3 Solution for Optimization considering average parking duration only Variable

x1 x2 x3
Z

= 0,75 68,3923
22,704 21,0222 112,1185

= 0,80 63,7555
23,5808 21,8341 109,1703

= 0,85 59,3562
24,4126 22,6043 106,3732

= 0,90 55,1768
25,2029 23,3361 103,7158

= 0,95 51,2011
25,9547 24,0321 101,1879

=1 47,4146
26,6707 24,6951 98,7804

From the solution in the Table 4.2, it shows that the higher the value of (level of satisfaction) the smaller the parking slot obtained especially the vehicle (motor cycle) having the highest average parking duration in comparison to the others. As average parking duration for all vehicle more balance (the differences were not very extreme), the resulting parking slot allocation also more balance between all three types of vehicle.

Proceedings of the 6th IMT-GT Conference on Mathematics, Statistics and its Applications (ICMSA2010) Universiti Tunku Abdul Rahman, Kuala Lumpur, Malaysia

471

4.3 Formulation Considering Both Parking Accumulation and Parking Duration


Optimization considering constraints for both average parking duration and parking accumulation with values of from 0.5 up to 0.7 was run using QM for Windows 2.0 with the solution is shown in the Table 4.4.
Variable

x1 x2 x3
Z

= 0,5 95,91958
17,49884 16,20263 129,621

= 0,55 89,74738
18,66595 17,28329 125,6966

= 0,60 83,93796
19,76446 18,30043 122,0028

= 0,65 78,46021
20,80025 19,25949 118,52

= 0,70 73,28654
21,77855 20,16532 115,2304

From the solution in the Table 4.4, it shows that the higher the value of (level of satisfaction) the smaller the parking slot obtained especially the vehicle (motor cycle) having the highest average parking accumulation and parking duration in comparison to the others. In contrast, parking slots for both Car and Becak increased with the increase of (level of satisfaction). Comparison of all three formulations suggests that the formulation considering parking accumulation only is the best option if the total number of optimal parking slot is used as a performance measurement. The formulation considering average parking duration only is clearly less preferable as it give the result of less number of parking slot and, in practice, it also usually does not relate significantly with the customer satisfaction.

5. Conclusions
The results of this study suggest that the formulation of optimization for parking slot allocation give significantly different results when the formulation consider various aspects of parking requirements and demands. Trying to fulfill all parking demands and requirements may lead to the optimization problem without any feasible region in the optimization. The most important aspects should come into the formulation before considering other aspects to see the realistic optimal result for the problem.

Proceedings of the 6th IMT-GT Conference on Mathematics, Statistics and its Applications (ICMSA2010) Universiti Tunku Abdul Rahman, Kuala Lumpur, Malaysia

472

References
1. Bruglieri Maurizio, Colorni Alberto, Lue Alessandro, The Parking Warden Tour Problem Proceeding of CTW09, pp. 11-15, (2009) 2. Cannon Stephen, Rehman Shakeb, Mendez Alberto, Vo Vincent, Ordoez Maria Ana, and Singh Amit, Optimization of Resource Distribution in The George Mason University Parking System George Mason University Study Report, (2003). 3. Cordone Roberto, Ficarelli Federico and Righini Giovanni : Bounds and Solutions for Strategic, Tactical and Operational Ambulance Location, a Dipartimento di Tecnologie dellInformazione Universit`a degli Studi di Milano, Proceeding of CTW09, pp 180-185, (2009). 4. Djakfar Ludfi, Indriastuti K. Amelia dan Wicaksono Achmad: Relokasi Pelataran Parkir Mobil Barang Kota Kediri, Simposium VII FSTPT, Universitas Katolik Parahyangan, (2004). 5. Hobbs, F. D.,: Perencanaan dan Teknik Lalu Lintas, Cetakan I, Gajah Mada University Press, Yogyakarta, (1995). 6. Rapp Matthias and Albrecht Christian : Capacity and Slot Management for Heavy Goods Vehicle Traffic Across the Swiss Alps, Paper presented at the 10th World Congress on ITS, Madrid, (2003). 7. Renta Iskandar, Jinca M. Yamin and Parung Herman : Standar Kebutuhan Ruang Parkir pada Rumah Sakit di Makasar, Simposium VII FSTPT, Universitas Katolik Parahyangan, (2004) 8. Setiawan Rudi : Penerapan Manjemen Transportasi Kampus Sebagai Upaya Mengurangi Penggunaan Mobil (Studi Kasus : Universitas Kristen Petra), Simposium VII FSTPT, Universitas Katolik Parahyangan, (2004). 9. Silva de Amal, Bus Driver Duty Optimization by Combining Constraint Programming and Linear Programming, ILOG (S) Pte Ltd. Singapore, (2000). 10. Sinaga Roberth and Priyanto Sigit : Penanganan Parkir Kendaraan pada Terminal dan Pasar (Studi Kasus : Kawasan Condong, Sleman dan DIY), Simposium VII FSTPT, Universitas Katolik Parahyangan, (2004). 11. Yosritzal, Gunawan H. and Justiawan A.A. : Analisis Perparkiran di Pasar Padang Raya, Simposium VII FSTPT, Universitas Katolik Parahyangan, (2004). 12. Widanengsih and Elkhasnet: Karakteristik Parkir di Lingkungan Universitas Bandung, Simposium VII FSTPT, Universitas Katolik Parahyangan, (2004).

Вам также может понравиться