RESEARCH ON EVALUATION OF LOCATION PLANNING FOR URBAN PUBLIC SERVICE FACILITIES BASED ON GIS SPATIAL ANALYSIS

Xianjuan KONG 1, Anrong DANG 1,2, Gongli LI 1 （1. Beijing Tsinghua Urban Planning and Design Institute, Beijing, 100084, P.R.C)

（2. School of Architecture, Tsinghua University, Beijing, 100084, P.R.C)

Abstract: Following the extensive urbanization, more and more

people swarm into cities. Demands for urban infrastructure and

land use, therefore, will change rapidly and drastically, and most

urban planning should be regulated or remade. Under the new

situation, location of public service facility should be paid more

attention to, while the general software platform could not assist it

well. Therefore, a new kind of traffic network system is proposed,

which is composed of both vehicle traffic model and pedestrian

traffic model, and corresponding algorithms are built or rebuilt.

Moreover, the evaluation indices are proposed to evaluate different

location planning scheme. Finally, new models and methods are

used in location analysis for urban public service facilities in case

cities, and show the effectiveness.

Keywords: Urban Planning, Public Service Facility, GIS,

Network Analysis

1. Introduction By the year 2009, the urbanization rate in China reached

46.6% (from National Bureau of Statistics of China), which was predicted to 70% by the year 2050. It means that a great deal of population will swarm into cities. Demands for urban infrastructure and land use, therefore, will change rapidly and drastically, and most urban planning should be regulated or remade. Urban public service facilities are closely related to daily life of citizen, the location planning of which should be paid more attention to.

Many foreign cities has gone through the urbanization, whose experience show that GIS (Geographic Information System) is the most suitable method to assist planning and designing for its spatial analysis function and spatial data management. It is since 1960s that GIS was used in urban planning and more popular in 1970s, from when researchers

and planners do a series of research and application [1]-[8]. However, most of the researches and applications are limited to raster analysis, simple accessibility analysis or simple service area analysis, whose result return only one optimal route, and could not to deal with more complicated situation and lead to large deviation from the reality.

Above all, a new kind of traffic network system, which is composed of both vehicle traffic and pedestrian traffic, is proposed to deal with location planning at different spatial scale. Moreover, new network analysis methods based on the new traffic system are rebuilt. Finally, evaluation indices are proposed to compare the location planning schemes quantitatively. And some application instances are introduced to verify that the new models and methods could manage the complex analysis of location planning. 2. Models

2.1 Vehicle network model

Fig.1 Composition of vehicle network

Vehicle model is composed of road section and intersection, showed as Fig. 1(a). There are some attributes for road section, such as length, limited speed, average speed, cost, and so on. Vehicle is delayed when crossing intersection, and it is different of going straight, turning left, and turning right. The influence factors of delay include traffic signals, layout of intersection, road grade, and so on. Cost is the weight used in network analysis, the value of

Road Section 1

Road Section 2

Road Section 3

Road Section 4

Intersection

(a)

Cost of Turning Left Cost of Turning Right

Cost of Going Straight

Cost

Cost

Cost Cost

Cost Cost

Cost

Cost

(b)

4220978-1-4244-9566-5/10/$26.00 ©2010 IEEE IGARSS 2010

which may be speed, time, length or fee, etc. Road section with cost and flow direction with cost together make up the vehicle network, showed as Fig. 1(b).

2.2 Pedestrian network

Fig. 2 Composition of pedestrian network

Fig. 3 Generation of pedestrian network

Pedestrian network is composed of walkway and crossing facility (showed as Fig. 2). Crossing facility include zebra crossing, overpass, and subway, which has different attributes and cost. There are three methods to generate pedestrian traffic facilities. (1) In larger dimension exists municipal road, footway, and zebra crossing are created automatically according to road centerline. (2) In smaller dimension without municipal road, the footway is created automatically by some rules according to land parcels. (3) In some local range, the pedestrian traffic facilities could be created by interactive drawing. The final pedestrian traffic network is showed as Fig. 3. 3. Methods

3.1 Algorithm to get access point of parcels Traditional access point of parcel is the center point

of parcel, which leads to great deviation from reality, especially when the parcel is larger. The improved algorithm is to find inflection points of parcel, and projected the points to road sections or walkways, showed as Fig. 4. The projection points are the access points of the parcel, finding any one of which means arriving the parcel

when search route.

Fig. 4 Parcel and its access point in traffic network

3.2 Algorithms of service area Based on accessibility, there are two type algorithms of

service area. Type 1: Service radius is given, along traffic network,

starting from one facility point, the parcels within the given radius make up of the service area of this facility. The analysis results of this method will show three situations: 1) a parcel serviced by only one facility; 2) a parcel serviced by several facilities; 3) a parcel serviced by no facility.

Type 2: Along traffic network, the shortest route and its lowest total cost from a facility to all parcels are calculated and saved. Therefore, a parcel will own the same number of shortest routes and lowest total costs as that of facilities. The parcel is assigned to the service area of the facility whose lowest total cost is minimal. This method is a complete distribution method. The difference between the two algorithms is listed in table 1.

Table 1 Comparison of two types of service area

Service Radius Without Service OverlayedType 1 fixed,given yes yesType 2 variable,computed no no

4. Case study 4.1 Data preparation and analysis process

The fundamental data involve spatial information (road centerline, parcel, and facility point) and attribute information (land use code, facility class, facility scale, and standard). The analysis process is as follows.

Step 1, initialization, import the fundamental data; Step 2, construct traffic network, find the shortest route

from any public facility point to any residential parcel and calculate the total travel costs;

Step3, according to the algorithms of service area, get

4221

the service area of every facility, calculate the statics indices, and give the evaluation.

4.2 Results and discussion The first case is welfare facility location in City A,

which is a new town planning. Present extent occupies a very small component of the planning extent, so sufficient land use could be offered to adjust the scheme.

Fig. 5 Type 1 service area analysis of city A

Location rules of welfare facility are setting up different class facilities according to different rank districts. If there is a higher class facility in a district, the lower class similar facility will not be set up any more, Showed as Fig. 5. The analysis of this location scheme is suitable to use vehicle network and type 1 service area method. From the graphic analysis result showed in Fig. 5, this location scheme is acceptable. The service radius of city class facility (No 2) is longer than any other facilities, whose service area is larger. Service area of the four facilities occupied all the study area and overlapped for some degree, which make the burden of every facility more rational. At the same time, the area and population added up by all the service area exceed that of the scheme, showed in table 2.

Table 2 Data of type 1service area analysis

Radius(m)

Area of R-LandUse

Ratio ofR-LandUse

Popula-tion

Ratio ofPopulation

1 2500 7956428.1 29.50% 280067 31.33%2 5000 18982921.5 70.40% 639594 71.50%3 2500 5917846.9 21.95% 186580 20.87%4 2500 3814270.8 14.15% 117471 13.14%

Total - 24692211.8 98.30% 873793 97.70%

Known from the analysis results above, type 1 service area method is suitable for location analysis in large scale.

The second case is location planning of primary school, which is the part of construction detailed planning in built-up area of City B. It is more suitable to employ

pedestrian network and type 2 service area method. According to section 2.2, the pedestrian network is built up based on the road centerline, showed in Fig. 6.

Fig. 6 Centerline network and pedestrian network The analysis results are showed by Fig.7. Based on Fig.

7, most distances between parcels and its service facilities are less than 500m, which is the national standard for pupils walking to school. However, there exist some irrational problems that utility ratio for some schools are too high while that of others too low. For example, utility ratio of school No 1 and No 2 are both less than 0.5, as well as the average service distance of which are less than 400m. It means that the school resource is waste. At the same time, utility ratio of school No 15 is too high, which mean supply of school is not adequate to the demand. Ratio of utility for school No 13 is only 0.2 (the red row showed in table 3), the same as that of the nearby schools No 9 and No 12. School No 9 is a present primary school, while school No 12 and No 13 are the new built primary schools, the location of which should be adjusted. On one hand, it can improve the utility ratio of the schools; on the other hand, the pupils in service area of school No 15 could be shared. The aim is to make the location of schools more balance. 5. Conclusions

The models and methods proposed in this paper have been used in location planning for public service facility in case cities, and obtained better application effect.

Due to limited space, the indices of population land use and distance are the key analysis indices in this paper. Moreover, more influence factors could be considered, such as economy, administrative district, facility scale, etc. And new service area algorithm could be built to analyze panning scheme from more aspects so that the final planning scheme is more satisfying and rational.

There is great practical significance for the new traffic

4222

model system composed of vehicle traffic and pedestrian traffic. In this paper, location of facilities with different class are analyzed by the two traffic network respectively, as well as the two traffic network could be combined to deal with more complicated analysis. For example, the accessibility between facility and parcel or two parcels in

bus station planning could be analyzed by combining the two networks, the travel time of which include walking time and time in the bus. Finally, it could be used to analyze accessibility between two cities in a region with multiple traffic modes, which offer a new method for regional planning.

Fig. 7 Type 2 service area analysis of City B

Table 3 Data abridged of type 2 service area

School Min Pupil Max Pupil Pupil Floor Area(m2) Utility Population Average1 735 1124 546 15294.3 19117.9 0.5 8285 387.32 788 1206 466 6153.5 20511.5 0.4 7071 381.2

… … … … … … … … …9 442 677 298 6907 11511.8 0.4 4526 370.1

12 654 1000 331 13606.4 17008 0.3 5020 52413 919 1406 286 19135 23918.8 0.2 4338 477.315 517 790 1242 10755.6 13444.5 1.6 18815 555.2… … … … … … … … …

Acknowledgements

This study was supported by National Natural Science Foundation of China (50678088), and National Project of Scientific and Technical Supporting Programs Funded by the Ministry of Science & Technology of China (NO.2006BAJ14B08)

Reference [1] C Sutcliffe, J Board, P Cheshire, “Goal programming and

allocating children to secondary schools in Reading,”

Journal of the Operational Research Society, 1984, 35 (8),

719–739

[2] S Benjamin, L C Richard, “A new network representation of

a “classic” school districting problem,” Socio-Economic

Planning Sciences, 1991, 25(3): 189~197

[3] A Yeh and M Chow, “An Integrated GIS and

Location-allocation Approach to Public Facilities

Planning-an Example of Open Space Planning,” Comput.,

Environ. and Urban Systems, 1996,20: 339~350,

[4] R L Church, “Location modeling and GIS. In Longley P A,

Goodchild M F, Maguire D J and Rhind D W, eds”,

Geographical Information System, Volume 1, Principles and

Technical Issues, 2nd edition. New York: John Wiley and

Sons, 1999, 293~303.

[5] T, Arentze A Borgers, H Timmermans, “A knowledge-based

system for developing retail location strategies,” Computers,

Environment and Urban Systems, 2000, 24: 489~508

[6] F Caro, T Shirabe, M Guignard and A Weintraub, “School

Redistricting: embedding GIS tools integer programming,”

Journal of the Operational Research Society, 2004, 55(8):

836-849

[7] K Oh, S Jeong., “Assessing the spatial distribution of urban

parks using GIS.,”Landscape and Urban Planning, 2007,

82(1-2): 25~32.

[8] A Comber, C Brunsdon, E Green, “Using a GIS-based

network analysis to determine urban greenspace accessibility

for different ethnic and religious groups,” Landscape and

Urban Planning, 2008, 86(1):103~114.

4223