System Dynamics Model for the Sustainable Development of
Science City
Yufeng Ho
Graduate School of Architecture and Urban Design,
Chaoyang University of Technology, Taichung, Taiwan
Email:hyfarch@ms32.hinet.net
ShuSong Wang
Department of Architecture, Chungkuo Inslitute of Technology, Taipei, Taiwan
Email:edge0711@seed.net.tw
Abstract
Hsinchu Science Park is an example of developing high technology industry in Taiwan.
Its rapid growth has resulted in the tremendous economic benefits. However, it also has
produced social and environmental impacts, such as traffic jams, environmental pollution
and heightened differences in social class. These impacts result in conflict between economic
growth and sustainable development of the science city. The purposes of this study therefore
are (1) to understand the interactive relationships between subsystems and indicators in the
sustainable development system of a science city, (2) to discuss the social, economic and
environmental issues occurring in this system, and (3) to device reasonable strategies for the
operation and management of a science city. The indicators for sustainable development of
a science city were selected by the method of Fuzzy Delphi with respect to the economic,
social and environmental aspects in the theory of sustainable development. A system
dynamics model was established by STELLA programming language to simulate the different
development scenarios of Hsinchu Science City. Our analysis reveals that the development of
the science city should comply with the objectives of (1) maximizing the profits from
industries in the science park, (2) minimizing the damages incurred by the science park on its
mother city, (3) decrease in consumption of natural resources, (4) reduction in environmental
pollution, and (5) attention to relevant social problems. In addition, an integrated
management strategy with due emphasis on economic, environmental and social aspects to
attain the sustainable development with economic prosperity, environmental and ecological
conservation as well as social equality
Keywords: system dynamics model, sustainable development, science city, Scenario
Introduction
Ever since its establishment in 1980, Hsinchu Science Park (HSP) has seen tremendous
development, bringing in remarkable economic benefits. Its annual average output between
1998 and 2000 accounted for 8% of Taiwan’s GNP, and its production doubled that of other
domestic industries. However, its rapid growth also resulted in adverse social and
environmental impacts including rising land price in Hsinchu, chaotic traffic, environmental
pollution and social segregation in Hsinchu City (HC). These problems all violate the 3E
(Environment, Economic and social Equality) principles of sustainable urban development.
Hence, efficient handling of the problems related to the growth of the HSP will be an
important indicator of whether Taiwan would succeed in developing into a green silicon
island in the 21* century.
The purpose of this study is to simulate the strategies of sustainable development for
HSC. (Fig.1) Urban development are closely related to social and cultural custom, ecological
and economic resources. The achievement of the formulation of sustainable development
strategy is require a systematic and flexible methodology; including an indicators system and
most importantly, a simulation model to relate the social, ecological and economic aspects of
urban planning together as a whole.
Fig 1. Hsinchu Science City
The application of system dynamics in urban systems has been raised by Forrester
since 1974. In recent years, considerable interest has been generated in the use of system
dynamics, as an aid in urban planning to assist planner in understanding the complex
interaction among social, ecological » economic and other factors. A number of people,
Nijkamp & Perrels (1994) » Gulen & Berkoz (1996) + Ho et al (2002), have been involved in
research in this field, resulting in a rapid growth in the literature on the subject. So it is
possible to search for the strategies of sustainable development in a system dynamics model.
Different urban development scenarios can be simulated with the model to shed light
on the effectiveness of sustainable development strategies and their impacts. The results
obtained in this study can serve as useful references for managing the HSC and devising
future development strategies with due emphasis on the economic, environment and social
aspects.
Research Method
Sustainable development is a dynamic process which recognizes the needs of everyone;
effective protection of the environment; prudent use of natural resources; and maintenance of
high and stable levels of economic growth and employment. In this case, the strategies to
be formulated will be more complex. It is obvious that the need has thus arisen for a new
and more capable method which will tend to produce improvement of system behavior to the
problem.
In HC considerable concern exists about urban problems, such as the deterioration of
living conditions, overcrowding and empty dwellings, traffic congestion, the shortage of open
space and many other aspects of urban social, economic and environmental situations. Most
of the existing planning methods in this context addressed only one or two of these three
aspects. This study is to develop an integrated framework for establishing a sustainable
urban structure to maintain a balanced relationship between human needs and urban
environment.
System dynamics can assist in strategy assessment and provides insights into possible
changes in the system during policy implementation (Sterman, 2000). A simulation model,
combining urban system analysis with system dynamics techniques is suggested. Through
model simulation, the proposed strategies are taken as changes in parameters and structure.
The objectives of sustainable development can be pursued to achieve a better quality of life
for every citizen, now and for generation to come.
Model Formulation
Formulation of Sustainability Indicators and Feedback Loop Diagram
Reviewing the establishment of Hsinchu Science City and its industrial development
and aiming to achieve more efficient management, we formulate science park industry +
population + housing + environmental pollution and economic five sub-systems (Appendix I)
and 52 indicators, as shown in Table 1. Fuzzy Delphi Method is employed for screening the
indicators. From the industrial sector, the government and the academia, 20 experts and
scholars were selected to assess the 52 indicators formulated in order to establish the
sustainability indicators for HSC.
threshold will be selected while those below will be screened.
indicators are obtained as marked with * in Table 1.
Table | Sustainable indicators and their assessment value
Indicators with experts’ assessment value above the
In the end, 22 sustainability
Subsystem Indicator Value {Subsystem Indicator Value
Number of companies 0.66 s_ [Rate of unoccupied houses 0.70*
a =}
Z | Productivity of SP staff o7m2* | 2 2, Average housing price 0.70*
s Zé
2 | Industry value 0.77% | % = [Rate of land development in SP 0.69
B @ 2
3 [Net revenue of capital 0.69 5B [Land area per SP staff 0.70*
E [Amount of imports 0.64 © |House rental rate in SP 0.58
& [Amount of exports 0.65 [Amount of dust fallen 0.69
R&D expenditure 0.70%* F [otal amount of suspended particulates _|0.72
‘Total population 0.63 $ [Daily sewage disposal per capita 0.83*
Population growth rate 0.74% & [Daily refuse production per capita 0.81%
&
Natural increase rate 0.56 'g [Amount of refuse collected per day 0.75
Social increase rate 0.70* = [No.of motorcycles per 1000 persons __ [0.64
Average size of household 0.63 2 _ |No. of vehicles per 1000 persons 0.70
J [Urban-to-total population ratio 0.65 2 —_|Number of factories registered 0.58
5 [Population density 0.76* Number of environmental _pollution|
g 0.82
5 lawsuits
va
Z [Age structure 0.71% [Amount of saving per household 0.68 *
3 | Education level 0.71% ‘Total regular income per family 0.71
Water consumption per capita 0.56 Housing-to-total family expenditure ratio 0.67
Power consumption per capita 0.57 | _|Rate of self-owned houses 0.61
B
Population of SP staff 0.59 g _ [No.of automobiles per 1000 persons _|0.55
$
Age structure of SP staff 0.60 2 |Rate of unemployment 0.70*
Education level of SP staff 0.71% | “% [Low income-to-total population ratio {0.57
>
[Area of agricultural land 0.46 2 No. of industrial units 0.61
2 =
2 2 [Urban-to-total area ratio 0.58 $ _|Industrial population 0.60
Za
‘Zo [Urban area per capita 0.72 Industrial-to-total population ratio 0.65
g 8
= & Population served by piped water 0.65 Industry value 10.82
© [Residential floor area per capita 0.70* [Area of industrial land 0.64
Note: * Adopted by higher values than the threshold : 0.695(HSP industry subsystem), 0.675(Population subsystem),
0.669(Housing/ Landuse subsystem), 0.760(Environmental pollution subsystem), 0.674(Urban economy subsystem)
Taking into consideration the development characteristics of HSC and linking the five
subsystems through variables such as industrial production, population and extent of
pollution.
development system of Hsinchu Science City were formulated and shown in Figure 2.
can be seen, there are 16 causal loops, 11 input and 5 information feedback ones.
4
The causal feedback loops of the different variables in the sustainable
As
3 Urban economy subsystem SPO .
* Amount of saving per household ae
* Total regular income per family ty
f 777] * Rate of unemployment \
‘Environmeptél — | * Industry value ‘
: ai wy Resources \
development
HSP industry subsystem
* Productivity of SP staff
*
Environmental pollution] \ 77775 * oy valle
subsystem expenditure |
* Daily sewage disposal per
capita
* Daily refuse production
per capita
* Number of environmental
pollution lawsuits
Housing/Landi ibsyste
Fopulation Subsyste lousing/Landuse subsystem’
+ Popul: h rate * Urban area per capita
a roe ntion: growth rats: * Residential floor area per capita
Environmental |* Social increase rate * Rate of unoccupied houses
management __[* Population density * Average housing price
: ~~.,]* Age structure
quality ‘-
* Rate of land development in SP,
* Education level * Land area per SP staff
ducation level of SP,
——* Input [_ 5 Producer CJ Consumer
— > Information CI Storage [] Others
feedback
Fig. 2 Integrated system model of the sustainable development of Hsinchu Science City
Model Formulation and Simulation Analysis
In view of the many parameters of an urban sustainable development system and their
complex relationships, we assume in this study that the behavior in such system is in a
continuous state in order to facilitate model formulation and simulation analysis. Therefore,
changes in system behavior are not a matter of probability; rather they are the results of some
causal loops of the variables. The formulated urban sustainable development model covers
Hsinchu Science Park and its mother city; historical data of the area from 1986 to 2000 are
taken as the basis to simulate and forecast the future development of HSC.
Figure 3 shows the dynamic model of the sustainable development system of HSC
which is written in the STELLA programming language (HPS, 2000).
As can be seen, the
model comprises 80 variables and 90 equations.
Among these equations, 10 are level
equations, 10 are initial value equations, 18 are rate equations, 32 ancillary equations, 15
constant equations and 5 graphic equations. With the system dynamic model formulated,
we conducted computer simulation and analysis of the HSC sustainable development
strategies.
AYQ d9UdI9g NYSUISH JOJ JUaUIdO]OAdp s]qQeLUTEISNs Jo JOPoUT WaysXs poyeISoIUI Oy} JO WeIseIP WTTALS ‘€ BIT
Justification of the Model
The model was justified with the historical data of Hsinchu City from 1986-2000. The
results were very close the real-world behavior. (Fig 4)
Total industry
= ‘| value (NT$ billion)
Za _ { Total amount of
a =~ = refuse (tons)
oe | Total amount of
—_— — —| sewage (1000 m3)
_ youl number of
= —~) residential units
= aaa (100 units)
— = Total number of
an === ~—| households
=— = er (100 households)
oe Total population
La eee (1000'persons)
Industry value of
fe SP
Z (NTS billion)
c No. of workers in
Z 7 SP
—— a (100 workers)
— i . .
me Historical data
PA --- Simulated value
1986 1990 1995 2000
Fig 4. Simulated data with respective to the historical data of Hsinchu City
Sensitivity Analysis
This study used the ‘Effect-Efficiency Matrix’ to select the sensitive variables in the
model (Kano, 2003). The matrix classified them into an active set consisting of 15 ancillary
variables and a passive set comprising 13 variables identified by the experts. During the test,
the value of each active variable was increased by 10% to produce 15 different values. Each
increment will result in a corresponding change in the value of the 13 passive variables. In
consequence, 196 curves were generated.
Because the changes in value could be positive or active, the values were normalized
7
by the arithmetic mean and the standard deviation. The relations between the active and
passive variables were divided into four classes of different scores as follows: (1) 0 for No
effect, (2) 1 for weak effect (lower than standard deviation), (3) 2 for median effect (within
standard deviation), and (4) 3 for strong effect (higher than standard deviation). The results
are shown in Table 2.
Table 2. ' Effect-Efficiency Matrix , of variables of sustainable development for Hsinchu
Science City
Indicator variables
SPSP|SPTIV| TP | NPI | HN |HUN|UHUN| RI | WPI | IV_| SIV | TIV | AS
3 3 2 2 2 2 2 2 2 2 2 2 26
3 3 2 2 2 3 3 2 2 2, 2 2 28
3 3 2, 2 2 2 2 2 2 2 2 2 26
0 0 0 0 0 0 0 0 2 2 0 2 6
> ILP 0 0 0 0 0 0 0 0 2 3 0 3 8
8 SILR. 0 0 0 0 0 0 0 0 0 0 2 2 4
= SILP 0 0 0 0 0 0 0 0 0 0 3 2 5
2 BR 0 0 2 3 2 2 0 2 2 2 2 2 19
s DR 0 i) 2: 1 2 2 0) 2 2 2 2 2 17
Ff HNM 0 0 0 0 1 1 0 i) 0 0 0 0 2
z HUNM 0 0 0 0 0 2 2 0 0 0 0 0 4
° PRC 0 0 0 0 0 0 0 3 0 0 0 0 3
WCPC 0 0 0 0 0 0 0 0 3 0 0 0 3
IWC 0 0 0 0 0 0 0 0 3 0 0 0 3
PS 9 9 10 | 10 ul 14 9 13 | 20 15: 15: 19 =
Note : SPSP : SP_staff_population SPLPM : SP_labour_productivity_modulus
SPTIV : SP_total_industry_value RDER : R & D_expenditure_ratio
TP : total_population ILR : industry_labour_ratio
NPI : net population_increase ILP : industry_labour_productivity
HN : households_number
HUN : housing_units_number
UHUN : unoccupied_housing_units_number
+ service_industry_labour_ratio
¢_industry_labour_productivity
RI : refuse_increase DR : death_rate
WPI : water_pollution_increase HNM : households_number_modulus
IV : industry_value HUNM : housing_units_number_modulus
SIV : service_industry_value PRC : refuse_per_capita
TIV : total_industry_value WCPC : water_consumption_per_capita
AS : Active Solution IWC : industry_water_consumption
SPSM : SP_staff_modulus PS : Passive Solution
The results indicate the most sensitive ancillary variables in the active set are SP
personnel modulus (SPSM), SP labour productivity modulus (SPLPM) and R & D
expenditures ratio (RDER) and the second sensitive variables are birth rate (BR) and dearth
rate (DR). In addition, the more is the degree of sensitivity, the higher does the degree of the
impact complexity it shows. .For instance, the most sensitive variable in the passive set is
water pollution increase (WPI). It then includes total industry value (TIV) ~ service
industry value (SIV) and industry value (IV).
Simulation of Development Strategies for Hsinchu Science City
Our simulation analysis shows that HSC has been and will be experiencing continuous
development from 1986 to 2021 in terms of SP industry value, total population, rate of
unoccupied houses, amount of sewage disposed, amount of refuse collected and total industry
value. Detailed simulated results of the policies are discussed in the following:
Scenario 1: economic growth strategy
Being the source of prosperity of HSC, the HSP should be given due priority in
strategy formulation. In other words, it is important to maintain a flourishing SP for the
economic benefit of its mother city. Without innovation and ongoing development, there is
no guarantee for lasting success of today’s hi-technology industries, which may then fall into
the fate of unsustainable traditional industries and become idle or oblivious. Retarded
growth of the SP will be a severe blow to the development of HSC. Our first scenario
simulation focuses on maintaining prosperity and sustainable growth of the SP. Possible
strategies for attaining such goal will include enhancing staff productivity and increasing
R&D expenditure ratio to foster growth in industry value.
According to historical data, the productivity of SP staff peaked in 2000 at NT$ 9.04
million per worker; while the average productivity per worker in the past 16 years is NT$
4.91 million, showing an average growth of 10.97%. Owing to the worldwide economic
downturn in 2001, it suffered a decrease in that year and thereafter has maintained a steady
growth of 5% per year. With reference to the highest productivity attained in 2001, the
upper limit of annual staff productivity is set to be NT$ 100 million.
Regarding R&D expenditure ratio, there has been a constant, though slow, growth,
from 4% of the total sales revenue to around 6%. The highest percentage of 7% was found
in 1998 and the average value was 5%. We assume that the R&D expenditure ratio will
maintain a steady increase and reaches 9% of total sales revenue in 2016.
Simulations under this scenario with and without strategy implementation were
performed and the results are displayed in Figs. 5 to 10. As can be seen, marked increase in
industry value is observed starting from 2006, five years after the policies have been
launched; the value is forecasted to rise more than two folds by 2021. On the other hand,
total population and total industry value also reveal growth though to a smaller extent. All
these indicate that rising industry value does foster the prosperity of the region, yet not
without costs. The rate of unoccupied houses is predicted to soar to 20%, 2% more than the
rate estimated without strategy implementation. This implies that the strategies will
indirectly lead to excess supply of houses. At the same time, both sewage disposal and
amount of refuse collected have shown increase, indicating that overdevelopment in industry
will aggravate the problem of environmental pollution.
Efforts made on hi-technology development will certainly bring about economic
benefits, though at the expense of further deterioration of the environment and disequilibrium
in housing supply. In other words, over-emphasis on economic development will upset the
social and environmental balance. Hence, to achieve sustainable development of HSC, the
industrial development of the SP should not be without constraints and due attention should
also be paid to the environmental and social issues in the area concerned.
SP industry value 2: SP industry value 2 1: population 2: population 2
090000 0
1000000
| | | 1
si
mal a
— : — Ls eet |
Al eee
f a
——
cao - e000 co} . ! . ‘
oo 206 aon 26 awk oor 7 aa a6 ml
Fig 5. Industrial value of SP Fig 6. Total population
‘:mocoplilileer nl Ss mocam ome i 1: water pollution increase 2: water potion increase 2
250000 00
p21
—_—_—"
im 178000 co mses
——
a
aus : i ' 100000 20-4
Son Ss an = ai oor 2 vn m6 al
Fig 7. Rate of unoccupied houses Fig 8. Total amount of sewage disposed
1s total industry value 2 total industry value 2
450000 00
. a mae
225.00. L . eis ee ae
127
st
250.004 ase 00-4
2001 2006 201 2016 02) 2001 2006 2011 2016 2021
Fig 9. Total amount of refuse collected Fig 10. Total industrial value of HC
Scenario 2: environmental protection strategy
The theory of sustainable development has always given ecological conservation the
top priority. With focus on protecting the environment, appropriate industrial and economic
development should be fostered without incurring negative impact on the ecology and at
minimum depletion of natural resources. To achieve such goal in the HSC, possible
strategies would involve restricting industrial and economic development through monitoring
the productivity and population of SP staff.
As mentioned in Scenario 1, global recession in 2001 has caused a reduction in
productivity. In our model, we assume that such declining trend will continue in future and
there will be a gradual decrease in productivity of SP staff from its peak in 2000 to an annual
10
productivity of NT$ 6 million per person in 2011 and thereafter it will remain unchanged.
At the same time, the model also sets the ratio of R&D expenditure constant, with neither
increase nor decrease.
As seen in the historical data, the total number of SP staff has increased from 8,275 in
1986 to 96,293 in 2001, showing an average annual rise of 18%. However, 2001 saw for
the first time negative growth, with a decrease of 349 employees compared with the total
staff population of 2000. In view of the declining trend in the SP staff population, the
model does not introduce any change in the parameter setting.
Figures 11 to 16 show the simulated results with and without policies implemented to
maintain environmental tolerance. Significant decrease in industry value is observed
starting from 2006. The declining trend becomes steadier after 2016. The forecasted
industry value of 2021 with strategy implementation is around 30% less than that without,
while there is little difference observed with respect to total population and total industry
value. That is to say, policies related to environmental conservation have little impact on
these two aspects. However, the rate of unoccupied houses has become better controlled,
showing a value of 16%, 2% less than the rate estimated without strategy implementation.
As for the amount of sewage disposed and amount of refuse collected, a slight decline can be
seen from 2006. Although the magnitude of decrease is not great, the quality of the
environment is on the whole improving.
1: SP industry value 2: SP industry value 2 1: population 2: population 2
\eocc 0 sueeee 00
a
——lT _2aee
ont ses
woe. i sxesco 00 Is
a a |
ta HN
00000 00-4 20000000
001 Fy mi wie co] 00 m6 a aie ami
Fig 11. Industrial value of SP Fig 12. Total population
eee semeeuatpaerens evn pliion eae 2: water polation increase 2
3 20000020
a a ee a ee
oar. ee ee 15000004 cen
— ea
js
ort ‘0000000
001 06 ma m6 m 001 2006 Eat 2016 at
Fig 13. Rate of unoccupied houses Fig 14. Total amount of sewage disposed
1 refs increase 2: efse inerease2 1c total industry value 2 total industy value?
00.80 «6000000
sd
Se ga Se
eee ayeeeeren
475.00. ee i 0000 00: eae
1 ent
ia
210.00 200 00-4
2001 706 aii m6 22 200 706 aw m6 2021
Fig 15. Total amount of refuse collected Fig 16. Total industrial value of HC
Aiming to conserve the environment and maintain social equality, the policies of
constraining industrial development in the SP do achieve reduction in rate of unoccupied
houses, amount of sewage disposed and amount of refused collected while showing little
impact on the total industry value in the mother city. Nevertheless, the forecasted hi-tech
industry value of 2021 will decline by around 30%, a significant economic impact worthy of
concern. Hence, too much stress on the environment and social equality but at the expense
of economic loss will still be a blow to urban sustainability; particularly so for a science city
whose main wealth draws from the industrial prosperity of the science park. Constraints
should be posed but steady development has to be preserved while industrial decline should
be avoided. In short, strategy-makers have to strike a balance between the 3E concerns so
as to realize sustainable development. That is to say, the policies implemented should target
at fostering industrial development, reducing environmental pollution and maintaining social
equality at the same time.
Scenario 3: Integrated management strategy
As seen in Scenarios | and 2, emphasis on any single aspect, be it economic growth or
environmental protection, cannot ensure equal progress in environment, economy and social
equality. Too much stress on increasing industry value will result in deterioration of the
environment and rise in rate of unoccupied houses. Although restricting the increase in
industry value may solve such problems, it is at the expense of maintaining steady growth in
the SP, which itself is also the goal of sustainable development of the HSC. Hence, a more
integrated approach should be considered. Possible strategies would involve both the SP
industries and the mother city. With respect to the former, the total population of SP staff
should be controlled and the R&D expenditure ratio should be increased; while in the latter,
measures should be taken to encourage home purchase so as to reduce rate of unoccupied
houses, promote recycling and reuse to cut down refuse production, and to foster economy of
water consumption.
As mentioned in Scenario 2, there has been an average annual increase of 18% in SP
staff beginning from 1986, which peaked at around 97 million in 2000. Owing to the
12
limitation of space in the SP, the total staff population will not rise indefinitely. Hence, the
study model sets the highest population of SP staff at 140 million. When the simulated SP
staff population reaches this limit, the model will stop further increase.
The parameter setting for R&D expenditure ratio in Scenario 1 is adopted for
simulation of Scenario 3. That is, there is constant steady growth in R&D expenditure ratio
with no upper limit. By 2001, the R&D expenditure ratio is forecast to reach 10% of the
total sales revenue of the SP.
Statistics shows that the average size of a household in HSC is 4.46 persons and is on
the decline at 2% each year. By 2000, the average size has dropped to 3.28 persons. With
the policy of encouraging home purchase, the rate of unoccupied houses has been on a steady
decrease at 2% per year and will reach a constant rate at 2011. Historical data indicate that
the daily average water consumption per capita has increased from 310 liter in 1986 to 330
liter in 2000; while the daily average amount of refuse collected per capita has decreased
from 1.05 kg in 1986 to 0.99 kg in 2000. Economy of water usage is encouraged in the
hope to reach the target daily average of 250 liter per capita, as stipulated in the National
Conference on Soil and Water Resources. Moreover, the falling trend of daily average
amount of refuse collected is to be maintained to reach the target daily average of 0.96 kg per
capita in 2016, which will remain unchanged thereafter.
Figures 17 to 22 displays the simulated results with and without policies of integrated
management implemented. As can be seen, these policies bring about a slight increment in
industry value. It is interesting to note that the population of SP staff does not rise along
with the increasing industry value; rather it is dropping a little. This indicates that the
policies of restricting total SP staff population and raising R&D expenditure have fulfilled
the purpose of adding value to the industry. There is hardly any change in the total
population and total industry value. Measures encouraging home purchase serve to bring
the rate of unoccupied houses down to 17%, 1% less than the rate estimated without strategy
implementation. Obvious reduction can be seen with respect to amount of sewage disposed.
This together with the decease in amount of refuse collected evidences the success of the
policies in enhancing the environmental quality of the HSC.
1 SP industry value 2: SP industry value2 1: population 2: population 2
200000 00 00000 00
' ee : :
ee eo
eo ——oo re
0000 004 <eon0 004
2001 76 ait a6 aml oor ave air aie mal
Fig 17. Industrial value of SP Fig 18. Total population
1: water pollution increase 2: water pollution increase 2
1: unoccupied house rte 2: unoccupied house rate
0.19 2000000
; cuales}
—s
wy 1 sao. 00 $n
al a
ose ‘e000 00-4
2001
2001 2006 20 2016 2m yl 2006 2011 2016 2021
Fig 19. Rate of unoccupied houses Fig 20. Total amount of sewage disposed
italics 1s total industry value 2 total industry value 2
7 ae ee |
00 0: ay oS all Geen (sna ce
ea see
2001
. 260000 00
2006 201 2016 202 2001 2006 2 26 2021
Fig 21. Total amount of refuse collected Fig 22. Total industrial value of HC
In short, integrated management policies implemented in the SP have achieved increase
in industry value, decrease in rate of unoccupied houses and reduction in pollution to the
mother city. All 3E concerns have been addressed with balanced emphasis paid on each,
thus serving the purpose of maintaining sustainable urban development. From this, we can
see that the sustainability of the HSC can be achieved not only by preserving continuous
industrial development in SP, but also with good integrated management of the mother city
from a macro perspective. Past experience has shown that economic development and
environmental conservation are often in conflict with each other. It seems inevitable that
fostering one will make the other suffer. However, our analysis above seems to indicate a
way out. As seen in our simulated development scenarios, while promoting economic
prosperity by adding value to the industry, the impact on the environment can be minimized
by appropriate control over industrial landuse and staff population. At the same time,
polices related to environmental conservation can help reduce resource consumption and
depletion as well as strengthen pollution control. Hence, integrated approach to
management with equal emphasis on the 3E aspects is the key to the success in achieving
sustainable urban development.
Conclusions & Suggestions
Conclusions
According to the simulation analysis, the following polices are possible means of
14
realizing sustainability in HSC:
1. Economic strategies
To promote industry development in the HSP, the emphasis should be on quality rather
than quantity. Increase in production quantity will imply larger labour population,
which may incur extra burden on the environment and cause social problems. Hence,
more effort and resources should be devoted to research and development, targeting for
more value-added outputs. In this way, economic prosperity can be realized through
increase in industry value with minimum negative impact on the mother city.
2. Environmental strategies
Measures aiming to reduce resource consumption and depletion should be implemented
to put pollution under control. Education and promotion on proper use of water and 3R
(recycle, reuse and reduce) should be launched. Facilities for refuse disposal and
sewage treatment should be added to enhance the efficiency in reducing sources of
pollution and curtailing further deterioration of the environment.
3. Social strategies
Science park employees are usually of higher education and higher pay, thus causing
social segregation. Real estate prices will tend to go up, which is unfair to other
residents of the mother city. Higher profits will attract developers to construct more
houses, leading to higher rate of unoccupied houses. In view of these problems,
sustainable development policies should aim at controlling the total staff population of
the HSP to avoid excessive increase. To reduce the rate of unoccupied houses, home
purchase promotion schemes should be launched providing greater benefits and
low-interest government loans to attract buyers.
Suggestions
Suggestions for future research are as follows:
1. HSC is a steady system with ongoing changes. Among the subsystems, there exist
feedback loops for control and adjustments to be made with respect to further system
growth. Hence, to realize the goal of attaining sustainable development, the assessment
and valuation of the sustainability indicators would yield reasonable standard value so
that proper limits can be set and appropriate amendments can be made. In this way, the
model formulated would bear closer resemblance to the reality, making the simulated
results more reflective of the future scenarios and possible changes.
2. Future studies can explore more in-depth micro-observation of a single subsystem.
Take for example the HSP industry subsystem; a more detailed and complete model of
the subsystem can be formulated by taking into consideration the impact of the four most
important industries, namely opto-electronics, semi-conductor, IC design, as well as
15
telecommunications and biotechnology, on the development of the HSP. It is also of
interest to investigate further into the interrelationships between two subsystems, such as
the HSP industry subsystem and the environmental pollution subsystem, to understand
better their causal links so as to establish more complete feedback loops between them.
3. Our simulated results reveal that environmental pollution can be effectively decreased at
the cost of reducing 30% of the industry value (NT$ 4000 billion). It is worth to explore
the cost of implementing other pollution control measures. Comparison between the
costs involved can shed light on which is a more economic approach to enhancing
environmental quality.
References
1. Forrester, J.W., 1974, Systems Analysis as a Tool for urban planning, in Nathaniel J. Mass,
Readings in Urban Dynamics: Vol.1, pp.13-28.
2.Gulen, C., & Berkoz, L., 1996, Dynamic Behavior of the City Center in Istanbul, Comput.,
Environ and Urban System, Vol.20 No.3.
3.Ho, Y.F., Wang, H.L & Lu, C.H., 2002, The Dynamic Simulation Model for the
Sustainable Development of Taichung City, Journal of Architecture, pp.107-128, Taiwan.
4. HPS, 2000, STEILA and STELLA Research: An Introduction to Systems Thinking, High
Performance Systems, USA.
5. Kano, N., 2003, Quality Control Story, JUSE, pp.46-48.
6. Nijkamp, P., & Perrels, A., 1994, Sustainable cities in Europe, Earthscan Publication.
7.Sterman, J., 2000, Business Dynamics: Systems Thinking and Modeling for a Complex
World, McGraw-Hill, pp.41-81.
16
Appendix I. STELLA diagrams of the subsystem of Hsinchu science city
R&D expenditure
SP industry vplue growth rate
SP staff modulus
$P labour productivity modulus
Fig23. STELLA diagram of science park industry subsystem
population density
land area
e
birth rate death rate
net population total floating population
floating population modulus
Fig 24. STELLA diagram of population subsystem
17
households number
unoccupied houses rate
using units constructed number
housing price rate housing units increase modulus
Fig 25. STELLA diagram of housing/landuse subsystem
water pollution
watér pollution in water pillurion decreas
sewage per bapita se disposal max
total industry sewage 2
sewage disposal modulus
total water eénsumphion per capita
industry sewage dispor
pon Seamer ee SP indugtry sewage
Selene pes copa moras, SP industry sewage modulus
total industry water consymptio
stn sy pci total SP industry water con3umption
industry sewage disposed modglus
water consumption per capit: SP industry water expense
industry water consumption industrial land area $P industrial area
refuse pollution
re la
refuse increase
refuse decreaa
total refuse disposal amount
refuse per capita
collectors number
collection rate
Fig 26. STELLA diagram of environmental pollution subsystem
18
total industry‘value
total industry value decrease
ry walue increase
service pdustry value
OO
sepvice industry it ervice industry value decrease
industry labour productivity ingustry labour population
Iabour ratio surviow indistry labour productive = tee Moe popu
industry value growth rate Service industry labour ratio service industry labors modulus
Fig 27. STELLA diagram of urban economy subsystem
19
Appendix II. System Dynamics Model of Sustainable Development for
Hsinchu Science City
1. SP industry subsystem
SP_industry_value(t) = SP_industry_value(t - dt) + (SP_industry_value_increase - SP_industry_value_decrease)
* dt
INIT SP_industry_value = 17043
INFLOWS:
SP_industry_value_increase = SP_staff*SP_staff_productivity
OUTFLOWS:
SP_industry_value_decrease = SP_industry_value
SP_staff(t) = SP_staff(t - dt) + (SP_staff_increase) * dt
INIT SP_staff= 8275
INFLOWS:
SP_staff_increase = SP_ staff *SP_ staff _increase_rate
R&D_expenditure = SP_industry_value_increase*R&D_expenditure_ratio
R&D_expenditure_rate = SMTH3(0.065,10,0.04)
SP_industry_value_growth_rate = (DERIVN(SP_industry_value,1))/(DELAY(SP_industry_value,1))
SP_staff_increase_rate = SP_industry_value_growth_rate*SP_staff_modulus
SP_ staff_modulus = 0.64
SP_ staff_productivity = SMTH3(7,6,3.5)*(SP_staff_productivity_ modulus*R&D_expenditure_ratio+1)
SP_ staff productivity_modulus = 2.7
2. Population subsystem
population(t) = population(t - dt) + (population_increase + floating _population - population_decrease) * dt
INIT population = 306088
INFLOWS:
population_increase = birth_rate*population
floating_population = floating_population_modulus*total_floating_population
OUTFLOWS:
population_decrease = death_rate*population
birth_rate = SMTH3(0.013,6,0.016)
death_rate = 0.0054
land_area = 104.0964
net_addition_population = population_increase+floating_population-population_decrease
population_density = population/land_area
total_floating population = 25510
floting_population_modulus = GRAPH(SP_industry_value_growth_rate)
(-1.00, -0.24), (-0.8, -0.21), (-0.6, -0.15), (-0.4, -0.09), (-0.2, -0.03), (-5.55e-017, 0.00), (0.2, 0.03), (0.4, 0.09),
(0.6, 0.15), (0.8, 0.21), (1.00, 0.24)
3. Housing/Landuse subsystem
household(t) = household(t - dt) + (household_increase) * dt
INIT household = 68587
INFLOWS:
household_increase = net_addition_population / household_modulus
house_unit(t) = house_unit(t - dt) + (house_unit_increase) * dt
INIT house_unit = 80061
INFLOWS:
20
housing_units_increase = household_increase*house_unit_modulus+house_construction_unit* housing_units
housing_units_increase_modulus*house_construction_rate
households_modulus = SMTH3(1.4,6,2)
house_construction_unit = 850
house_unit_modulus = 1
unoccupied_house_rate = unoccupied_house_unit/house_unit
unoccupied_house_unit = house_unit-household
house_construction_rate = GRAPH(house_price_rate)
(0.9, 0.7), (1.31, 0.755), (1.72, 0.821), (2.13, 0.914), (2.54, 1.05), (2.95, 1.27), (3.36, 1.55), (3.77, 1.77), (4.18,
1.89), (4.59, 1.96), (5.00, 2.00)
house_price_rate = GRAPH(SP_staff_increase)
(0.00, 0.9), (2000, 1.06), (4000, 1.27), (6000, 1.43), (8000, 1.68), (10000, 1.99), (12000, 2.38), (14000, 2.81),
(16000, 3.30), (18000, 3.93), (20000, 5.00)
house_unit_increase_modulus = GRAPH(SP_staff_increase)
(-2000, 0.00), (-1600, 0.2), (-1200, 0.4), (-800, 0.6), (-400, 0.8), (0.00, 1.00), (400, 1.00), (800, 1.00), (1200,
1.00), (1600, 1.00), (2000, 1.00)
4. Environment pollution subsystem
refuse_pollution(t) = refuse_pollution(t - dt) + (refuse_increase - refuse_decrease) * dt
INIT refuse_pollution = 0
INFLOWS:
refuse_increase = refuse_amount
OUTFLOWS:
refuse_decrease = refuse_disposal_amount
water_pollution(t) = water_pollution(t - dt) + (water_pollution_increase - water_pollurion_decrease) * dt
INIT water_pollution = 0
INFLOWS:
water_pollution_increase = person_sewage+total_industry_sewage
OUTFLOWS:
water_pollurion_decrease = disposal_modulus*max_disposal_of_sewage
disposal_modulus = 0.9
industry_land_area = SMTH3(250,5,120)
industry_sewage = total_industry_water_expense*industry_sewage_modulus
industry_sewage_modulus = 0.6+industry_value_growth_rate
indutry_water_expense = 210
max_dispose_of_sewage = 121400
personal_refuse = 0.00105
person_sewage = person_sewage_modulus*total_person_water_expense
person_sewage_modulus = 0.8
person_water_expense = SMTH3(0.33,10,0.29)
purge_population = (population+SP_staff*0.5)*purge_rate
purge_rate = 1
refuse_amount = purge_population*personal_refuse
refuse_disposal_amount = 900
SP_industry_area = 625
SP_industry_sewage = total_SP_industry_water_expense*SP_industry_sewage_modulus
SP_industry_sewage_modulus = SP_industry_value_growth_rate*0.01
SP_industry_water_expense = 180
total_industry_sewage = SP_industry_sewagetindustry_sewage
total_industry_water_expense = industry_land_area*indutry_water_expense
21
total_person_water_expense = water_supply_population*person_water_expense
total_SP_industry_water_expense = SP_industry_area*SP_industry_water_expense
water_supply_population = population+SP_staff*0.5
5. Urban economy subsystem
industry_value(t) = industry_value(t - dt) + (industry_value_increase - industry_decrease) * dt
INIT industry_value = 41768
INFLOWS:
industry_value_increase = industry_labors*industry_labor_productivity
OUTFLOWS:
industry_decrease = industry_value
service_industry_value(t) = service_industry_value(t - dt) + (service_industry_value_increase -
service_industry_value_decrease) * dt
INIT service_industry_value = 9372
INFLOWS:
service_industry_value_increase = service_industry_labour_productivity*service_industry_labours
OUTFLOWS:
service_industry_value_decrease = service_industry_value
total_indutry_value(t) = total_indutry_value(t - dt) + (total_industry_value_increase -
total_industry_value_decrease) * dt
INIT total_indutry_value = 51141
INFLOWS:
total_industry_value_increase = industry_value+service_industry_value
OUTFLOWS:
total_industry_value_decrease = total_indutry_value
industry_labors = population*industry_labor_rate
industry_labor_productivity = SMTH3(4.1,8,0.65)
industry_labor_rate = SMTH3(0.18,6,0.21)
industry_value_growth_rate = (DERIVN(industry_value,1))/(DELAY(industry_value,1))
service_industry_labors = population*service_industry_labor_rate*service_industry_labors_modulus
service_industry_labor_productivity = SMTH3(1.2,9,0.4)
service_industry_labor_rate = SMTH3(0.13,6,0.08)
service_industry_labors_modulus = GRAPH(SP_industry_value_growth_rate)
(0.00, 1.00), (0.1, 1.01), (0.2, 1.02), (0.3, 1.04), (0.4, 1.06), (0.5, 1.10), (0.6, 1.14), (0.7, 1.16), (0.8, 1.18), (0.9,
1.19), (1, 1.20)
22
CC BY-NC-SA 4.0