MULTIPURPOSE SIMULATION SYSTEMS FOR REGIONAL
DEVELOPMENT FORECASTING
The planning for public sector development of the national economy should
take into account main socio-economic peculiarities of regions. Otherwise, the
realization of each alternative of the development of national economy in a
particular region can unpredictably affect its social-economic conditions. In the
process of development of public goods production infrastructure, social and
economic conditions can change as a result of close interaction and their constant
mutual interdependence.
Nowadays, the responsibility of provision of public goods is shared between
different levels of Ukrainian government. However, “the Center” is. still
responsible for fair distribution of resources to make sure that all the needs of
population are satisfied. The method for regional development planning, we want
to discuss in this paper, is referred to as a large system that enables one to use a
system approach to scientific solution of arising problems. Here stands the point
where a problem of rational planning becomes a problem of optimal system
control. The efficiency of public service delivery can be measured in economic and
social terms and may be estimated with the help of modern information
technologies, widely known economic-mathematical methods, and computer
engineering tools.
In the countries of former Soviet Union and other post-socialist countries the
problem of control of social-economic development for the regional planning is of
current interest of great importance, because of multidimensional way of
development of their economies in transition period. The significant role of
planning for socio-economic development makes it possible to manage a wide
range of controlled parameters to achieve the optimal result. On the one hand, the
latest circumstances favor improvement in planning efficiency, but, on the other
hand, make this problem more complicated elevating it to the multidimensional
level.
Increased significance of this issue for national economy requires explicit
solution based on the present-day scientific methods. Regional automated control
systems are developed, constructed and operated in our country to improve control
at the territorial level. As a rule, subsystems of social sphere development control
are included into regional Automatic Management Systems (AMS) to take into
account social-economic factors of social sphere development, studying
peculiarities of interaction between various model components. Along with
account-informational problems these subsystems allow for finding a solution for
forecasting and optimization problems. The solution of such problems in the AMS
framework should provide a variety of the most efficient alternatives of regional
development scenarios based on fundamental social and economic logic.
The solution of certain optimization and forecasting problems for regions has
been a standard practice for domestic and foreign researches during several recent
years. In particular, several approaches to the mentioned problems of such
character have already been introduced into software of the operating AMS.
Methods for formalization of problem statements, applied here, vary greatly from
one to another in their targets, as well as in various structural and functional
features. In spite of the benefits that derive from solutions of such problems one
should remember that their diversity does not allow for using all advantages of
modern mathematical methods and computer technologies. This, as well, does not
give the possibility to take into account a sufficient degree of various factors in
their interrelations, i.e. to realize and apply a system approach to the problem
solving. It would be advisable to construct, on the basis of a unified mathematical
methodology and general technical and information bases, a system of forecast-
optimization problems which will help to find substantiated answers to a broad
class of questions on regional development planning. The work, which is presently
carried on in this direction under the supervision of the National Academy of
Sciences of Ukraine is one of possible ways for resolving of these complex issues.
For resolving of forecast-optimization problems at the regional level it is
necessary to take into consideration various arising aspects - demographic, purely
2
social, economic, productive, financial, organizational, judicial, etc., which are
closely interconnected. These characteristics hamper the use of such traditional,
mathematical methods as mathematical programming, critical-path approach, and
theory of random processes. In fact, the method of computer simulation processes
is the only mathematical method suitable for reflecting various features of systems
for social-economic development at the regional level which is capable of giving
the required accuracy of results. This method may be successfully used due to its
considering of random factors’ behavior when obtaining not only local (individual)
but also integral (generalized) estimates. This feature characterizes a spread of
figures being forecasted and checks the accuracy of the obtained forecast.
Among various known simulation methods the most suitable one is the
method of automaton simulation [1,2] developed at the Institute of Cybernetics of
the National Academy of Sciences of Ukraine. This method possesses essential
practical advantages and is presently rigorously substantiated from the
mathematical point of view and proved successful in solving a great number of
practical problems mainly in the fields of transportation economics, inventory
control, and rational design.
A multidimensional vector having the property of Markoff processes is used
as a basis for construction of a large system automaton model. If this process has a
special property called the property of conditional independence of components,
each component of the process may be identified with a relevant state of particular
automaton from the standpoint of automata theory. If this takes place, the signals
circulating between automata reflect interrelation of system characteristics of
certain automata. The existence of this property permits executing the operation of
simulation for different states of a system, component-by-component. In /1,2/ it is
shown that under fairly broad range of assumptions with respect to the initial
Markoff vector (a finite number of interrelated random actions) the property of
conditional independence of its components may be satisfied. A general practical
method of realizing such a procedure is described below.
The essence of the approach for forecast-optimization problems of regional
development is not to reduce the problem task to a simple prediction of the future
condition. On the contrary, such approach is a complex algorithm that envisages an
active reproduction of a free process, multistage classification for possible control,
selection of the relevant choices, and estimation of process characteristics. That is
why each forecast-optimization problem, at its final stage, leads to the construction
of a model of the complex system with many branches intended for solving a
whole aggregate of individual interrelated problems.
For constructed and operated models to meet the requirements placed upon
them, the leading algorithms should make provisions for periodic interference of a
researcher in the model operation process. The researcher’s interference in the
model operation chiefly amounts for the problems of a considered class, for
execution of the following two functions:
1) classification, selection, change, and correction of control alternatives;
2) choice and assignment of the modes for model operation, definition of a
list of output figures. According to this function, the algorithm of the
model for forecast-optimization problem possessed the general structure
shown at the following block-diagram in Fig.1.
Selection of data files used for solving problems
v
Selection, change and correction of control options
v
Choice and assignment of modes for model operations
v
Simulation unit (automaton model)
v
Information accumulation
Vv
Unit for statistical execution of simulation process
v
Formation of output information
v
Fig. 1. Block-diagram of the general structure of the model of forecast-
optimization problem
The stated considerations are used as a basis for elaboration of social-
economic developments of forecast-optimization problems at the regional level.
The first stage of this approach includes several models which have, in designers’
opinion, the foremost significance as far as their applications are concerned. All
models of the first stage correspond to a general structure of problems for a
considered class, allow for the dialog-mode operation, and are constructed in the
form of simulative automaton models.
The analysis of real situations of social-economic development of a region
shows that the influence of demographic factors is in primary importance for
modeling process. Thus, the first constructed model of the first stage system
hierarchy is the model for forecasting the development of demographic situation in
a region. The model uses extensively all advantages and possibilities of the chosen
mathematical method. This model possesses a central place in the general structure
of a system, and its outcomes are successfully used for solving other similar
problems.
To obtain close forecast figures of population size with respect to gender and
age groups, it is necessary for the model to take into account the elements of
development of family relations of the considered population. The constructed
model, along with classification of the population with respect to gender and age
groups, also takes into consideration a number of marriages in terms of ages, and
simulates the process of marriages and divorces. Incoming and outgoing migration
flows are taken into consideration by estimating gender and age distributions,
marriages and divorcees of those who migrate.
One of the solid applications of this method has been a model of Forecasting
of the Regional Population (DEPROG), which is represented in the following
formulae where:
B,,(t) is the number of married couples in the region for women of age x and men
of age y at the end of year t (t = 0,/,..., T);
a,(t) (x=0,100) is the number of unmarried women of age x in the region at the
end of year t;
¢,(t) (y=0,100) is the number of unmarried men of age y in the region at the end
of year ft;
g,(t) (x=16,100) is the actual ratio of the number of marriages with women of age
x during the current year to the total number of unmarried women of age x;
d,(t) and d’,(t) (x=16,50) is the number of births in the region by married and
single women of age x respectively during the year ¢ + 7;
e,(t) (x=0,100) is the number of deaths of women of age x in the region during the
yeart + 1;
ft) (y=0,100) is the number of deaths of men of age y in the region during the
yeart + 1;
h,(t) (x=16,100) is the number of divorces of women of age x in the region during
the year t + J;
m,(t) (i=1,.) is the current annual in-migration flow of type 7 (total number of
immigrants of type / entering the region during the year t + /);
nt) (j=1,/) is the current annual out-migration flow of type j (total number of
emigrants of type J leaving the region during the year ¢ + /);
The table of conditional functional transitions (CFTT) is written as the
following system of independent difference stochastic equations:
b, (t+1)=(0,b,4 40+
. min {100 ,y+30 }
a,,(1)8,,(t)u}_, min Je, - 1(¢), os 21 OB (tus,
max {16 ,y-30
+ min {100 ,y+30 } #
a,,(t)8.4 (t)u;_.
max {16 ,y-30 }
ld fyi
+5 m (t)0 Vig Wi yee ~ Ox-ry-1 2) aa foo
7 cy, (t) + aoe
max {16 ,y-30
€ 1 (t)
~ Beat, ya min ‘ao (x 430} = hy (t)uy_, ~
a(t)+ Yb)
max {16 ,x-30 }
ie 16 ,x + 30 if x = 16 ,46
= De n,(1)0jV Wi yes x = 16 100 ;y =4x-30,x+ 30 if x = 47,70
x — 30,100 if x = 71,100
miteor9d BOLO ei
a(t+l)=mayia.,0+ >> Saga th +h +3 O0-v)a,-4,,@ 20, —-
mayl6x-3q, ae 0 Yb KC) il a,,()+ Degn
maxl@y-30 mal
min {100 ,»++30}
win 100 xa30 4-1) Bt (EM yy min fen (1), & a8. On|
3 - max {16,y-30
os) an WOOT * — (0 =o jay (= 1-100;
~— 4.18.1 (Ou. “
max {16,y-30}
min {100,430 } by i Mei
c, (t+ 1) = max {Pepa a x min {100,x+30 J
>, elt
max {16,y-30 } a aO+
ax {16,x-30
+
min {100 ,y+30 } L
DY hea Mey. + Ym OU - 0B, ~
ved
max {16,y-30 }
fy @)
~ c, (4) min {100 ,y+30 }
cy1(t) + Son
max {16 ,y-30 }
min {100,y+30 } i . ——— es
— min fe,.(0, a. (8.1 (OUt« Yn, Od- v’ )B pr = 1.100 );
i=l
max {16,y-30 }
a(t+=70-6)> [4()+ 4.0];
x=16
c(t+l=y-6 YS [awa];
min foo. r.30 }
di(t+1)= p,(t) 2! b,. Ly _(t) (x = 16,50);
max 316 ,x-30
min {100 ,x+30 }
e(t+N=gita n+ Y brays() (x= 1.100);
max {16,x-30 }
min {100 ,x+30 }
fUFD=q Mey (OF LY beryr() (xe = 1.100);
° max {16,x-30 }
g,(t+1)=&, (4) x = (16... 100 );
min {100,x+30 }
A(t+N)=n,@) Yb.4,10 x =16...100);
max {16,x-30 }
mj (t+1)= 6,(t) G= Lue f);
n,(t+1)= v(t) (§=1..7)
Initially, the model is designed to be used for forecasting on the basis of
appropriate data files related to the City of Kiev, Kiev Region, and some selected
towns and rural regions of Ukraine.
One of the factors profoundly effecting social-economic development of a
region is the development of production of private and public enterprises on its
territory. This factor links regional development to the branch planning of national
economy, its territorial realization. The production factor, as well, strongly
depends on present demographic and social conditions in the region and, in
particular, on availability and growth of a labor force. That is why the second
model of the first stage problems relates to the study of interrelations between
production development in a region and migration flows. The model uses statistical
data accumulated during several previous years. The second model is based on
applying the first model to a study of demographic development of an observed
region, as well as of the adjacent regions.
The third model of the first stage refers to optimization and forecasting of
development of service provided to population of the region. In the well-known
publications by domestic and foreign authors this problem is solved in a variety of
ways which are, in fact, are very general approaches to this problem.
During last few years in many towns of our country there has been an
observed tendency towards transforming these towns into exemplary ones. The
concept of an exemplary town up to the present time has only qualitative character
and comprises a big number of improvements from the viewpoint of economic
conditions, further gain in productivity of labor, progressing in the sphere of
service provision, and ecological conditions. In the future, all means aimed at
development of towns and big cities should be used with maximum efficiency for
improvement of the welfare of population. That is why the tendency of
transformation of towns into the exemplary ones should be taken into account
while solving forecast-optimization problems of regional development and, in
particular, the problem of optimization of the service delivery to population.
The problem arises at the level of development of mathematical formalization
of an exemplary town. In spite of the difficulties of working out such formalization
criterion, this problem seems to be solvable. In the model, the process of time
consumption by the regional population is simulated on the basis of certain
statistical data and the average time budget. The relation between the average re-
quired time and the average free time may be taken as an adequate indicator for
measuring efficiency of functioning of service sphere.
Main model dialog mode operation with the user friendly interface allows for
optimal distribution of allotted capital investments among various types of services
throughout the territory of a region. Such an alternative for development is given a
preference when one Ukrainian hrivna of expenditures yields the maximal increase
of the average free time of the regional population. The same index may be used to
obtain a qualitative estimate according to which the given town can be granted a
status of the exemplary town. Apparently, for towns of various categories (national,
regional and rayon, centers) a different grading should be used to determine
whether the town is considered to be an exemplary one. The index should
proportionally grow with the growth of the population well-being and socio-
economic development of a region.
The developed model may also be used to obtain recommendations about
rational distribution of resources among different towns and places for
organization of public service provision.
The designers of the model hope that their work will contribute to
improvement of a well-being of population and to the efficiency of capital
investment in big cities and towns in various parts of our country and primarily in
the City of Kyiv.
References
1. Bakaev A., Kostina, N. Yarovitskii N., Simulation Models in
Economics, Naukova Dumka, Kiev, Ukraine, (1978), In Russian.
2. Yarovitskii N., Kostina N., Probabilistic Automata and Simulation,
Cybernetics and Systems Analysis, No.3, pp.20-31, (1993), in Russian.
3. Nina Kostina, Automaton Modeling as an Instrument for the
Forecasting of Complex Economic Systems, 21-st International
conference of System Dynamics Society, New York City, USA, 2003.
10
4.
Alekseev A.A., Kostina N.I., Kononets A.Y., Financial-Economic
Expert Systems, Kiev, SCARBI Publishing, Ukraine, 2004, In
Ukrainian.
ll
Title:
MULTIPURPOSE SIMULATION SYSTEMS FOR REGIONAL
DEVELOPMENT FORECASTING
Authors:
Nina Kostina,
Doctor of Economics, Professor,
National Academy of Tax Service of State
Tax Administration of Ukraine
Address: 8 Feodory Pushinoy Str., Apt.222, Kiev 03 115, Ukraine
Tel. +38(044)452-7774
E-mail: kni40@yahoo.com,
Victor Bazylevich
Doctor of Economics, Professor
Kiev Taras Shevchenko National University of Ukraine
Address: 90a Vasilkovskaya St., Kyiv 03022, Ukraine
Tel./Fax +38(044)259-7008
E-mail: bazyl@econom.univ.kiev.ua
