Method for the proposal and evaluation of urban structures for compact cities using an expert system

Abstract

The world population is continually increasing, but the Japanese population is decreasing. By 2050, the Japanese population is calculated to be close to 100 million people, and by 2060, 40% of the Japanese are estimated to be aged 65 years or more. Data show an increase in the declining birth and aging rates of the population. The development of high-concentration urban structures will be needed to solve such problems. However, a detailed vision of the future is yet to be clarified and in the process of consideration. In this study, we developed an instrument called “Population Migration Tool,” which is based on plural planning and policy, after we formed a future population distribution model. In addition, We also established a high-concentration urban structure model with this tool with the aim of considering the kind of urban structure local cities should target in the future.

Keywords

Local city ; Compact city ; Future population ; Visualization ; Concentration urban structure

1. Introduction

1.1. Background and purpose of the study

While the world population is continually increasing, the Japanese population is decreasing. In 2015, the Japanese population calculated to be less than 100 million people (Table 1 ). In addition, Fig. 1 shows that in 2060, 40% of the people will be over 65 years old. The data show an increase in the declining birth and aging rates of the population. To solve such problems, high-concentration urban structures will need to be developed. Local governments are considering compact city projects, which could recover the vitality of the central district and establish a life base for the residents.

Table 1. Would and Japanese population (thousand).

	2010	2020	2030	2040	2050	2060
World	6,916,183	7,716,749	8,424,937	9,038,687	9,550,944	9,957,398
Japan	128,057	124,100	116,618	107,276	97,076	86,737

Fig. 1. Age structure of 2010 and 2060 in Japan.

In addition, local cities are aiming to develop high-concentration urban structures as part of their future vision. However, a detailed vision of the future has yet to be clarified and is in the process of consideration.

Therefore, in this study, we developed the “Population Migration Tool,” which is based on plural planning and policy of local government, after we formed the future population distribution model. Moreover, we formed a high-concentration urban structure model with this tool, and we aim to consider the type of urban structure local cities should target in the future.

1.2. Review of related studies

Several studies have focused on compact cities worldwide. For example, Ahn et al. (2014) proposed the direction of urban regeneration of small and medium-sized cities of the China. Xiaolu et al. (2013) verified carbon dioxide emissions. Boquet (2014) organized the features of Dijion City and considered the good points of the city. Jabareen (2006) proposed a Matrix of Sustainable Urban Form, which evaluates the sustainability of different urban forms. Prasanna and Nurul Handayani (2015) measured urban expansion process in Indonesia, and analyzed to understand temporal and spatial characteristics of urbanization. Schneider and Woodcock (2008) examined the similarities and differences in urban form and growth that have occurred across 25 mid-sized cities in North America and the United States.

Moreover, many studies have examined compact cities in Japan. Takahashi and Deguchi (2007) verified the cost of the formation effects. Uchida et al. (2009) and Kobayashi et al. (2010) verified carbon dioxide emissions. Yamane et al. (2007) proved the effects of consumer behavioral patterns. However, few studies have examined the formation method of a compact city in detail for a targeted local city in Japan.

1.3. Study methods

In this study, the target area is Takamatsu city, which is the capital of the Kagawa prefecture. At first, using 100-meter mesh data, we analyzed urban structures based on land use and population distribution. Next, using a primary factors cohort, we formed a future population distribution model with the 100-meter mesh data after we estimated the future population. In addition, we developed a “Population Migration Tool” according to plural plans and polices of the target area. Moreover, we migrated the population to the concentration base using the tool. Then, we simulated the future urban structures based on the use of high-concentration urban structures. Lastly, we evaluated the model.

1.4. Overview of the expert system

An expert system is a system designed to solve a professional problem. The proposed system consists of three interfaces comprising a knowledge base to store the knowledge of experts, an inference engine to calculate the information that wants of the user by using the knowledge in the knowledge base, and the interface that provides information in an easy-to-understand form to the user.

2. Summary of the target area

As previously mentioned, the target area is Takamatsu city, which is in the Kagawa prefecture. In this section, we organized the urban structure of the target area based on land use and population distribution.

2.1. Land use

The urban structure in the target area was analyzed using the state of the land use based on the 100-meter mesh data. The land use was classified into 11 categories, namely, “Rice Fields,” “Agricultural Land,” “Forest Areas,” “Waste Areas,” “Building Areas,” “Highway,” “Another Area,” “Rivers and Lakes,” “Beach Areas,” “Sea Areas,” and “Golf Fields.” Table 2 shows the number of the land use mesh. The meshes of “Rice Field” continually decreased, whereas the meshes of “Building Area” continually increased (Table 2 ).

Table 2. Number of land use meshes.

	1991	1997	2006	2009
Rice Fields	6108	5412	4704	3415
Agricultural Land	1362	1416	1101	1074
Forest Areas	2688	2609	2877	2988
Waste Areas	125	89	124	62
Building Areas	3252	3780	4894	6295
Highways and Railways	162	224	232	48
Another Area	471	643	442	555
Rivers and Lakes	761	749	621	599
Beach Areas	13	13	2	0
Sea Areas	98	98	40	0
Golf Fields	32	39	35	36
Total	15,072	15,072	15,072	15,072

2.2. Population distribution

The urban structure in the target area was analyzed using population distribution with the 100-meter mesh data. Population distribution was classified into seven categories of “0 people,” “1–20 people,” “21–40 people,” “41–100 people,” “101–200 people,” “201–300 people,” and “over 300 people.” Table 3 shows population by age group, and Table 4 shows the number of meshes of the population distribution. Fig. 2 shows the population distribution of 2010. The younger population between the ages of 0 and 14 shows a decreasing trend as demonstrated, and the older population of over 65 years old is on an increasing trend (Table 3 ).

Table 3. Population by age group.

	1995	2000	2005	2010
0–14	67,456	62,861	60,505	57,943
15–64	282,376	279,332	271,957	255,599
65 above	62,746	74,009	84,314	93,667
Uncertain age	48	478	1349	12,220
Total	412,626	416,680	418,125	419,429

Table 4. Number of meshes of the population distribution.

Population	Mesh	Population	Mesh
0	5609	101–200	645
1–20	4668	201–300	32
21–40	1992	301	6
41–100	2120	Total	15,072

Fig. 2. Population distribution of 2010.

2.3. Summary of the urban structure

We organized of the target area based on the land use and population distribution. The meshes of “Building Areas” greatly increased, and the urban structure widened. However, the younger population showed a decreasing trend, and the older population indicated an increasing trend. As a result, the low-density population increased. Therefore, the target areas needed to plan carefully the compact city.

3. Formation of the future population distribution

In this section, we calculated the future population by using the primary factor cohort and the population by age. In addition, we formed the future population distribution of a 100-meter mesh using the future population and the population distribution of the 100-meter mesh.

3.1. Calculation of future population

We calculated the future population for 2020, 2030, 2040, 2050, and 2060 by using the primary factor cohort. Table 5 shows the future population. As is the case with future Japanese population, the future population of the target area shows a decreasing trend. In addition, the younger population between the ages of 0 and 14 is on a decreasing trend, and the older population over 65 years old is on an increasing trend. Thus, the declining birth and aging rates of the population are progressing in the target area as well.

Table 5. Future population by age group.

	2010	2020	2030	2040	2050	2060
0-14	45,630	37,229	31,330	30,667	24,441	22,183
15-64	200,586	179,640	165,512	138,651	121,260	110,023
65-	71,760	89,194	91,374	97,126	95,378	84,916
Total	317,976	306,062	288,216	266,443	241,079	217,122

3.2. Formation and evaluation of future population

Next, we formed the 100-meter mesh future population distribution model based the future population of the city and the 100-meter mesh data. Table 6 shows the number of meshes of the future population. From these projections, the type of future urban structure was predicted to be low-density. For the 2010 urban structure, the category for “1–20 people” has 4668. For the 2060 urban structure projection, the population will be around 5691. In addition, in 2010, over 100 people recorded 683, while 2060 urban structure will have 248. From the above results, the population of the target area is lower than in the present situation, and concern is raised on the spread of urban structure towards low-density.

Table 6. Number of meshes of future population.

	2010	2020	2030	2040	2050	2060
0	5609	5616	5621	5639	5656	5691
1–20	4668	4861	5098	5451	5611	5734
21–40	1992	1925	1853	1812	1783	1771
41–100	2120	2053	1978	1803	1717	1628
101–200	645	584	493	346	287	233
201–300	32	26	23	19	16	13
301 above	6	7	6	2	2	2
Total	15,072	15,072	15,072	15,072	15,072	15,072

4. Formation and evaluation of high-concentration urban structure model

In this section, we formed the high-concentration urban structure model. First, we developed the population migration tool after we built a base relying on the plural master plans of the target area. Then, we simulated a high-concentration urban structure model based on the tool. Finally, we evaluated the model.

4.1. Summary of base

We built the base, the axis, and the five-area division based on the seven plans. This study used the following for references, “Basic Plan of Land Use in Kagawa Prefecture” (Kagawa Prefecture, 2011a ), “Creation Plan of Setouchi Garden City” (Kagawa Prefecture, 2011b ), “Basic Policy of Urban Planning toward the Realization of High-Concentration Urban Structure” (Kagawa Prefecture, 2007 ), “Takamatsu Regional Planning Master Plan” (Kagawa Prefecture, 2012 ), “Takamatsu Comprehensive Plan -The Fifth-” (Takamatsu City, 2008a ), “Takamatsu City Planning Master Plan” (Takamatsu City, 2008b ), and “Basic Plan of Beautiful Takamatsu City” (Takamatsu City, 2011 ).

4.2. Setting the population migration rules

Next, we set the population migration rules. Fig. 3 shows a flow chart of the rules.

• Designated as a non-inhabitable area with land use

If the land use of the mesh is the “Rice Areas,” “Agricultural Land,” “Forest Areas,” “Waste Areas,” “Rivers and Lakes,” “Beach Areas,” “Sea Areas,” and “Golf Field,” then the mesh is designated as a non-inhabitable area. In this case, the population in the meshes moved towards the other transmigration area.

• Designated as a non-inhabitable area with a five-area division

If the five-area division of the mesh is the “Forest Area,” “Natural Park Area,” and “Natural Protection Area,” then the mesh is designated as a non-inhabitable area. In this case, the population in the meshes moves to the other transmigration area.

• Population migration to urban area

If the “five-area division is an urban area” and the “base is a wide-area base” or the “axis is a wide-area axis,” the minimum population density of the mesh is set at 100 people/ha. Additionally, if the “five-area division is an urban area” and the “within the base” or “overlaps with axis,” then the minimum population density of the mesh is set at 60 people/ha.

• Population migration to agricultural area.

Fig. 3. Flowchart of the population migration tool.

If the “five-area division is an agricultural area” and the “base is a wide-area base” or the “axis is wide-area axis,” the minimum population density of the mesh is set at 60 people/ha. Additionally, if the “five-area division is an agricultural area” and “within the base” or “overlaps with axis,” then the minimum population density of the mesh is set at 40 people/ha.

4.3. Developing the population migration tool

Moreover, we developed a population migration tool based on the four population migration rules. The tool could change arbitrarily the rules and simulate the high-concentration of the urban structure. Fig. 4 shows a screenshot of the tool.(Fig. 5 ).

Fig. 4. Interface of the population migration tool.

Fig. 5. 2010 Concentration of urban structure model (left), and 2040 concentration of urban structure model (right).

In sheet 1, we set the year of the population distribution. Then, we designated the non-inhabitable area based on the land use and a five-area division.

In sheet 2, we set the base area based on the wide-area, local, and the community bases. We put the base area at a distance from each base point. Next, we set a minimum target population for the base area and axis.

In sheet 3, we migrated the population to the area in the five-area division in the urban area, and in sheet 4, the five-area division is the agricultural area. Moreover, we established five population concentration rules: 1) “The population moves to the other transmigration meshes,” 2) “Not migration population,” 3) “Keeping the 2010 population,” 4) “The minimum population density of the mesh set at XXX people/ha,” and 5) “The minimum population density of the mesh increases XX% of the population.”

In sheet 5, we migrated the rest of the population to the concentration base, and we formed the high-concentration urban structure model.

In the last sheet, we evaluated the high-concentration urban structure model based on the number of meshes and the base area. As a result, we simulated a high-concentration urban structure model using this tool. Then, we formed the 2010 and 2040 concentration urban structure models based on this tool.

4.4. Evaluation of concentration urban structure model

We evaluated the high-concentration urban structure models by using the population distribution and the distance from urban facilities.

4.4.1. Population density

Table 7 shows the high-concentration urban structure models based on population distribution. As shown in the number of meshes of the population distribution in the 2010 urban structure, “1–20 people” has 4668 (31.0%), and the 2040 urban structure has 5451 (36.2%). The 2010 high-concentration urban structure model has 297 (2.0%), and the 2040 high-concentration urban structure model has 405 (2.7%). In the 2010 urban structure, over 200 people have 38 (0.2%), and the 2040 urban structure has 21 (0.1%). The 2010 high-concentration urban structure model has 52 (0.3%), and the 2040 concentration urban structure model has 27 (0.2%).

Table 7. Number of meshes of future population.

	2010 urban structure	2010 concentration urban structure	2040 urban structure	2040 concentration urban structure
0	5,609 (37.2%)	10,925 (72.5%)	5,639 (37.4%)	10,925 (72.5%)
1-20	4,668 (31.0%)	297(2.0%)	5,451 (36.2%)	690 (2.7%)
21-40	1,992 (13.2%)	449(3.0%)	1,812 (12.0%)	636 (4.7%)
41-100	2,120 (14.1%)	2,412 (16.0%)	1,803 (12.0%)	1,790 (16.9%)
101-200	645(1.3%)	929(6.2%)	346 (2.3%)	1,008 (3.1%)
201-300	32(0.2%)	52(0.3%)	19 (0.1%)	21 (0.2%)
301-	6(0.0%)	8(0.0%)	2 (0.0%)	2 (0.0%)
Total	15,072 (100.0%)	15,072(100.0%)	15,072 (100.0%)	15,072 (100.0%)

4.4.2. Population by base

Table 8 shows the high-concentration urban structure models based on the population by base. The population of the urban area increased in comparison with before the concentration. Moreover, population of the wide-area and local base increased compared with before the concentration.

Table 8. Number of meshes of future population.

	2010 urban structure	2010 concentration urban structure	2040 urban structure	2040 concentration urban structure
Urban area	199,894.6(60.8%)	277,150.6(84.4%)	159,581.1 (61.8%)	217,277.7 (84.2%)
Agricultural area	100,622.4 (30.6%)	51,222.0(15.6%)	76,883.3(29.8%)	40,797.7 (15.8%)
Other area	27,855.6(8.5%)	0.0(0.0%)	21,611.0 (8.4%)	0.0(0.0%)
Wide-area base	48,526.6(14.8%)	64,444.1(19.6%)	35,625.0(13.8%)	49,921.3(19.3%)
Local base	171,976.6(52.4%)	230,093.7(70.1%)	140,415.5(54.4%)	181,215.2(70.2%)
Community base	8,602.8(2.6%)	4,511.8(1.4%)	6,907.9(2.7%)	3,498.1(1.4%)
Outside of base	29,322.9(30.2%)	99,266.5 (8.9%)	23,440.9(29.1%)	75,127.0(9.1%)
Total	328,372.6 (100.0%)	328,372.6(100.0%)	258,075.4(100.0%)	258,075.4(100.0%)

4.4.3. Population by distance from facilities

Table 9 shows the high-concentration urban structure models based on the population by distance from the train station. Table 10 shows the high-concentration urban structure models based on the population by distance from the general hospital.

Table 9. Population by distance from the train station.

	2010 urban structure	2010 concentration urban structure	2040 urban structure	2040 concentration urban structure
0-1,000	226,594.4	260,679.1	178,814.9	204,310.0
1,001-2,000	69,455.9	60,076.2	54,698.9	47,819.5
2,001-3,000	24,348.2	7,057.9	18,839.1	5,488.1
3,001-4,000	4,970.2	559.4	3,674.1	457.8
4,001-5,000	1,565.1	0.0	1,029.8	0.0
5,001-	1,438.8	0.0	1,018.5	0.0
Total	328,372.6	328,372.6	258,075.4	258,075.4

Table 10. Population by distance from the general hospital.

	2010 urban structure	2010 concentration urban structure	2040 urban structure	2040 concentration urban structure
0-1,000	20,328.7	22,341.1	16,258.8	17,745.6
1,001-2,000	57,990.7	72,481.8	46,735.8	57,120.0
2,001-3,000	95,215.2	102,908.5	75,476.8	80,966.5
3,001-4,000	85,147.9	72,808.8	67,136.5	57,458.4
4,001-5,000	51,665.7	47,712.1	39,152.8	37,040.7
5,001-	18,024.5	10,120.3	13,314.6	7,744.2
Total	328,372.6	328,372.6	258,075.4	258,075.4

The population in closer proximity to the train station increased, and the population at a greater distance from the train station decreased (Table 9 ). Meanwhile, the population in closer proximity to the general hospital increased, and the population at a greater distance from the general hospital decreased (Table 10 ).

Therefore, the high-concentration urban structure models in the target area showed a high-density population and traffic convenience increased.

5. Conclusion

In this study, we built a future population distribution model by using a primary factor cohort. In addition, we developed a population migration tool based on plural master plans. Finally, we formed and evaluated the high-concentration urban structure models using that tool. The results are as follows.

(1) The younger population is on a decreasing trend, and the older population is on an increasing trend in the target area.

(2) As a result of the formation of the future population structure of target area, the future population gradually decreased. Additionally, we predict that future urban structures will move towards low-density.

(3) We formed a high-concentration urban structure model using the population migration tool. The rules of the tool can be changed arbitrarily. For this reason, we were able to consider a variety of high-concentration urban structure models.

(4) The concentrated urban structure model indicated a high-density population and increased traffic.

This study makes possible to visualize the future urban structure. This study will consider policies needed in creating effective urban structures, which is our goal for the future.

References

1. Ahn et al., 2014 Ahn, J. G., Jeong, H.C., Lee, J.H., Byeong, H., 2014. A study on the plan indicator for the activation of urban regeneration in small and medium-sized cities. In: Proceedings of the 9th International Symposium on City Planning and Environmental Management in Asian Countries. pp. 89–93.
2. Boquet, 2014 Y. Boquet; Transportation and sustainable development in a mid-size french city, Dijon; Int. Rev. Spat. Plan. Sustain. Dev., 2 (2) (2014), pp. 52–63
3. Xiaolu et al., 2013 X. Gao, J. Ji, F. Wang; Spatial pattern of transportation carbon emission based on behavior zones: evidence from Beijing, China; Int. Rev. Spat. Plan. Sustain. Dev., 1 (2) (2013), pp. 3–14
4. Jabareen, 2006 Y.R. Jabareen; Sustainable urban forms: their typologies, models, and concepts; J. Plan. Educ. Res., 26 (2006), pp. 38–52
5. Kagawa Prefecture, 2011a Kagawa Prefecture, 2011a. Basic Plan of Land Use in Kagawa Prefecture.
6. Kagawa Prefecture, 2011b Kagawa Prefecture, 2011b. Creation Plan of Setouchi Garden City.
7. Kagawa Prefecture, 2007 Kagawa Prefecture, 2007. Basic Policy of Urban Planning toward the Realization of Concentration Urban Structure.
8. Kagawa Prefecture, 2012 Kagawa Prefecture, 2012. Takamatsu Regional Planning Master Plan.
9. Kobayashi et al., 2010 Kobayashi, M.Kobayashi, T.Ikaruga, S., 2010. Evaluation of compact city by applying the area division system. In: Proceedings of the 7th International Symposium on City Planning and Environmental Management in Asian Countries. pp. 1–8.
10. Prasanna and Nurul Handayani, 2015 D. Prasanna, K. Nurul Handayani; Measuring the urban expansion process of Yogyakarta city in Indonesia; Int. Rev. Spat. Plan. Sustain. Dev., 3 (4) (2015), pp. 52–63
11. Schneider and Woodcock, 2008 A. Schneider, C.E. Woodcock; Compact, dispersed, fragmented, extensive? A comparison of urban growth in twenty-five global cities using remotely sensed data, pattern metrics and census information; Urban Stud., 45 (3) (2008), pp. 659–692
12. Takahashi and Deguchi, 2007 M. Takahashi, A. Deguchi; Method for cost-benefit evaluation system of impact by forming compact city; J. City Plan. Inst. Jpn., 42 (3) (2007), pp. 487–492
13. Takamatsu, 2008a Takamatsu City, 2008a. Takamatsu Comprehensive Plan -The Fifth.
14. Takamatsu, 2008b Takamatsu City, 2008b. Takamatsu City Planning Master Plan.
15. Takamatsu, 2011 Takamatsu City, 2011. Basic Plan of beautiful Takamatsu City.
16. Uchida et al., 2009 G. Uchida, T. Ujihara, M. Taniguchi, S. Hashimoto; Can we realize the new rural structure with low environmental load? Focus in automobile CO-2 emissions; J. City Plan. Inst. Jpn., 44 (3) (2009), pp. 361–366
17. Yamane et al., 2007 K. Yamane, J. Zhang, A. Fujiwara; Influences of city compactness on travel behavior patterns in a local city: analysis using an aggregate-type discrete choice models incorporating inter-alternative similarities; J. City Plan. Inst. Jpn., 42 (3) (2007), pp. 595–600