You are on page 1of 14

STOCHASTIC MODELS FOR POLITICAL PROCESSES

Sahib Esa and Boyan Dimitrov

Kurdistan Regional Government, Erbil, IQ


sahib esa@yahoo.com

Department of Mathematics
Kettering University
Flint, MI 48504, USA
bdimitro@kettering.edu

Key Words: election model, governmental model, k-out-of-n systems, markovian models,
party model, political modeling, survival characteristics

ABSTRACT
We extend the Probabilistic modeling power from stochastic processes (queuing theory,
dams, and reliability) on creation of some realistic models for studying political formations,
and show ways to find their survival characteristics. Our focus is on a dynamic ”democracy”
model, where the times of survival (existence at the political scene, duration of stay in
power, in leading coalition, governing survivability, life time distribution, longevity, etc.)
can be derived from the model. Markovian models of spending time in certain sets of states
are explored. We use results from queuing theory and reliability to create models for political
units in democratic environment from various points of view: We propose a party model,
an election model and a governmental model Some discussion on statistical properties and
practical implementations of these models are briefly sketched. We are confident that other
political schemes also can be modeled using appropriate probabilistic tools.

1. INTRODUCTION
Modeling politics need clear visual concepts and understanding. Its inclusion into quan-
titative models is a challenging task. Politics has broad products an fields of expressions:
from dictatorships to democracy, from peaceful society through revolutions, from interna-
tional coalition and interstate competitions, from wars to collaboration and mutual help
and support, from planned moves and actions on the international arena towards reaching
hidden targets (like winning elections, stabilization of currency) in a particular country, or

1
globally. Politicians are not the brightest persons in mathematics. If a mathematician is
going in politics, usually this person is leaving the field of mathematics. Applications of
mathematics for politics have been challenged in a huge scale in war times: construction of
modern weapons, in breaking information codes, in creation of war strategies, in optimiza-
tion of numerous important processes. In times of peace mathematics is of important use
in computational social and economic sciences in social simulation in general, and for polit-
ical science in particular. However, interest of scientists in development of specific political
models does exist. Many political scientists attempted at least to discuss the possibility to
apply mathematics and statistics in studies of political processes. Approaches include wide
spectrum of mathematical and computational tools. Our brief review of available literature
is summarized in the next paragraph.
Ciofi-Revilla (2009) presents a computational model of a political system, called SimPol
to highlight the fundamental structures and processes of politics in a generic society. There
are used combined methodologies of object-based modeling (OOM), the Unified Modeling
Language (UML), and the methodology of research programs. SimPol demonstrates that
computational models of entire political systems are methodologically feasible and scientifi-
cally viable. According to the author, SimPol can also build on and progress beyond any the-
ory and research done in advance. It helps the understanding of how politics operate across a
variety of domains (simple vs. complex) and levels of analysis (local, national, international).
In a recent textbook Clark et al. (2008) explore the intersection of fuzzy mathematics and the
spatial modeling of preferences in political science. The authors outline the potential bene-
fits of a fuzzy approach to the study of ambiguous or uncertain preference profiles. Taylor’s
book (1995) is applying ideas in modern mathematics to problems in international conflict
resolution, using discrete and deductive mathematics. Models of escalation, game-theoretic
models of international conflict, political power of regulation, and social choice are discussed.
The quantitative methods are presented in strongly applied settings for political science and
the social sciences. Doyle (2000) shows that nonlinear methods can provide new insight into
existing political questions. His concept of political complexity explores what happens when
political actors operate in a dynamic and complex social environment. The author is devel-
oping rigorous nonlinear models capable of empirical verification. Approaches from the areas
of landscape theory, computational modeling, time series analysis, cross-sectional analysis,
dynamic game theory, duration models, neural networks, and hidden Markov models are
involved. Monroe (1997) provides an important evaluation of the current state of empirical
political theory and a valuable guide to future developments in political science. Ordeshook

2
(1986), Morrow (1994), and Hafer (2007) emphasize on the use of games theory in modeling
political processes. Formal political theory seeks to develop formal, mathematical models
of political and economic processes. These authors in their books and articles attempt to
integrate the modern developments in the theory of games (decision making with multiple,
interactive decision makers) as the basis for the synthesis. Topics covered include models of
elections and of committee processes, the demand and supply of public goods, and surveys
of game theory and social-choice theory.
In our article we use probability tools to model basic components in the big political
games. We are confident that theories applied in reliability and queuing (service) system
modeling provide a good assembly of approaches for analysis of various sides of the polit-
ical products (risks, and costs inclusive). A principal advantage of these theories is that
they admit their models to be simply and easy translated into political models. Then the
obtained intermediate results allow to determine immediately various characteristics in po-
litical activities. Here we present simple mathematical models of political party, elections,
government constructions, and discuss aggregated numeric characteristics related to their
existence. Some of these results have been partially reported by Esa and Dimitrov at the
MMR conferences in Moscow (2009), and in Beijing (2011). The presentations enjoyed warm
reception. In our opinion, these models can be used to dynamically control the products of
political processes by taking actions which would change the parameters of proposed models
in desired directions.

2. A POLITICAL PARTY MODEL


When model a political party first we take into account that this is a formation within
a population (society) of finite sources of active (eligible) individuals N . There are several
political parties competing for supporters and members from the non-party people. Each
individual may decide to join a party or not at any time. Chances may vary with the time but
at first we assume it stationary. However, the existence (formation) of a party requires certain
minimum of members, say M : A member can quit the party at any time. Independently,
individuals from the non-party population may join the party. Simultaneously, the pool of
other parties operates in a similar way, possibly with different intensities. We assume that
members of one are not allowed to join directly (defect) another party.
Locking at this mechanism and dynamics, the following natural mathematical model
to describe this process: There is a population of N eligible individuals (voters, the non-
party people of a country). The non-party population maces ”the source”for the party

3
populations.We label the non-party population as ”the zero party”. These are considered
as sources of calls. Each party offers N available servers (the memberships for a party we
assume that it can accommodate the entire nation), There are R existing parties. Therefore,
we look at this democratic party system as at a finite (closed) queuing model. There is
a source of N individuals - potential customers, and R service systems with N servers in
each, One individual from the source may feel the need to become member of the party i
with an itching intensity λi > 0, (i = 1, 2, . . . , R.) Once a sit in the party i is occupied, it
stays occupied for a certain non-zero time Si , which is the service time on the i-th group
of servers. After a call is served (i.e. the party member decides to quit), it gets back to
the source. The successful existence of a party is made if at least K sits are permanently
occupied. The life time of a party is represented by the duration of time while at least K
servers remain occupied in the service system i.
Queuing theory (a relatively good recent presentation can be found in Gross and Harris,
1998) provides approaches to such characteristics as in finite source closed service systems.
We use results from Obretenov and Dimitrov (1978) and support it with our additional
calculations for the considered specific case as a part of the political process. An additional
simplification comes from the assumption that the number of servers in each group (these
are the party seats) potentially equals to the number of calls in the sources (the number of
eligible individuals from the population).
Each party is considered as a service system of N available seats (servers) in each. The
input rate from the source towards the system j is λj > 0; j = 1, . . . , R. Only one call at
a time can be emanated from the source. This call always has a server to take it, as in the
”one-to-one” basis. Hence, we can apply a non-homogeneous Poisson input model for each
of the R service systems, and assume also exponentially distributed service times. At system
j the call spends some random time Sj , and then comes back to its source.
However, we prefer another model leading to similar results, but easy to adopt. Look
at any particular individual in this society. He has R + 1 options where to go. Imagine, he
goes fort and back on all the options with the intensities assumed to attract this individual
to a particular type of servers. And does the motion up it gets into a steady state. A simple
Markovian model with random walk of a particle on the R + 1 states j = 0, 1, 2, . . . , R
exponentially distributed service times Sj of parameters µj > 0, at state j and competing
exponentially distributed times of parameters λj while particle is at state 0 can be applied.
It shows that the stationary probabilities Pj that the call is in service at system j , or that

4
it is in the source (this is when j = 0), are given by the expressions
( ) ( )
λ1 λR λj λ1 λR
P0 = 1 + + ... + ; Pj = 1+ + ... + , j = 1, 2, . . . , R. (1.1)
µ1 µR µj µ1 µR

We assume individuals independently express their political preferences. Therefore, a simpli-


fied multinomial model describes the entire spectrum of the party’s life in the country, with N
independent active eligible individuals. The coordinates of the random vector (X0 , X1 , , XR )
represent the number of individuals occupying the sits in each subsystem. They are dis-
tributed according to the multinomial law
N!
P (X0 = k0 , X1 = k1 , . . . , XR = kR ) = P k0 P k1 . . . PRkR ,
k0 !k1 ! . . . kR ! 0 1
(1.2)
k0 + k1 + . . . + kR = N.

Hence, the chance of the j th party to exist (at a minimum Mj members required for this
purpose) is having Binomial probability P (Xj ≥ Mj ) = 1 − B(Mj ; N, Pj ) with the same Pj ,
given by (1.1). The probability generating function of the distribution (1.2) is given by the
expression
{ )
P (z0 , z1 , . . . , zR ) = E z0X0 , z1X1 , . . . , zRXR = (P0 z0 + P1 z1 + . . . + PR zR )N , |zj | ≤ 1. (1.3)

It allows particular calculation of various average characteristics, correlations etc. The use of
these analytical results with the introduced parameters in respective statistical data allows
estimation of these parameters and answers to various interesting statistical questions.
Remark 1. Further dynamic analysis will allow to determinin the duration of existence
of a party, and other probability characteristics related to its active life. Important param-
eters here are the intensities λj and µj , j = 1, . . . , R. We call the λj ”rates of sympathy
towards party j” in this society. The rates µj , ”rates of dissatisfaction from party j” in the
society. These are conditional names having nothing in common with political terminology,
and are related to process trends only.

2.1. Characteristics of the democratic political party model


First we use the results from the model of political parties in a country, with N indepen-
dent active, eligible free individuals, considered above. The multinomial model in equation
(1.3).describes the entire spectrum of the party’s life in the country The coordinates of the
random vector X ⃗ = (X0 , X1 , . . . , XR ) represent the number of current individual members

5
of each party. The probability generating function of their joint distribution from equation
(1.3). allows particular calculation of various average characteristics, correlations etc.
The average number of members in party j is E(Xj ) = N Pj , and its variance
equals V(Xj ) = N Pj (1 − Pj ) .
The correlation coefficients ρij between the counts Xi and Xj of the parties
labeled as i and j are given by the expressions
√ √
Pi Pj
ρij = Corr (Xi , Xj ) = − for i ̸= j, i, j = 0, 1, . . . , R.
1 − Pi 1 − Pj
The correlation is always negative, since parties compete for one and the same pool of
potential members. Interesting quantities are the correlation coefficients ρ0j of the parties
j = 1, . . . , R with the ”party” of the free individuals. Their absolute values may be used as
a kind of rating indices for the parties in the country. The more the absolute value value
|ρ0j | is, the higher the party j rating is. For the party j it is possibly interesting to find out
that party kj for which it is fulfilled

ρjkj = min (ρji ) .


i̸=j

This is the main rival to the party j on the political scene in this country. An interesting
observation here is, that the correlation coefficients between parties (as well as the rating)
do no depend on the total population size N .
Probability for party j to be dissolved must be considered under condition that
this party is active, i.e. given that Xj ≥ Mj + 1. This is the probability that someone of
its members will decide to quit the platform when the party members are at the critical
quantity Mj + 1. Thus, this probability equals

dj = P (Xj (t + ∆t) = Mj | Xj (t) = Mj + 1)


( )
N (Mj + 1) · µj ∆t
Pj j (1 − Pj )N −Mj −1
M +1
= .
Mj + 1 1 − B(Mj ; N, Pj )
Probability for party (platform) j not to exist (to be in a ”sleep” state). Each
party can be considered as an idea (platform) gathering people. There are Xj people in the
society gathered around the platform of party j, and Xj , j = 0, 1, . . . , R are determined
by (1.2). If party j requires minimum of Mj + 1 mates to stay alive, then the probability
that this party does not exist is given by the expression

P (Xj ≤ Mj ) = B(Mj ; N, P j) ,

6
with the Pj , determined by equation (1.1).
Further dynamic analysis will allow to determine the duration of existence of a party.
Let us note that when we look at the Markovian process of the changes in the platform’
(party’s) states, we imaginary associate with it an absorbing Markov chain whose absorbing
states are {0, 1, . . . , Mj }. The states {Mj + 1, Mj + 2, . . . , N } are transient. Then, the
duration of the existence of party j on the political scene equals to the time the above
absorbing Markov chain spends in its transient states.
Denote by Bk0 , k1 , ..., kR the average time the process spends at the transient set of states
(sojourn time), given that it starts its random walk at the state described in the left hand side
of equation (1.2). Taking into account the infinitesimal intensities of the process transitions,
the respective average spending times immediately before a transition, and the respective
sojourn times starting with the state after transition, we arrive to the following system of
equations for the sojourn times Bk0 , k1 , ..., kR , valid for kj > Mj + 1, for all j = 1, . . . , R :


R ∑
N
Bk0 ,k1 , ...,ki −1, ..., kR 1
Bk0 , k1 , ..., kR = ki µ i ∑R ∑R + ∑R ∑R
i=1 Ki =Mi +1 k0 j=1 λj + j=1 kj µj k0 j=1 λj + j=1 kj µj


R ∑
N
Bk0 +1, k1 , ...,ki , ..., kR
+ (k0 + 1) λi ∑ ∑R .
i=1 Ki =Mi +1 k0 Rj=1 λj + j=1 kj µj
(1.4)
On the boundary layers ki = Mi + 1 for some j = 1, . . . , R the equations (1.4) are still
valid with Bk0 , k1 , ...,ki −1, ..., kR = 0 Also is true that Bk0 , k1 , ..., kR = 0, if kj ≤ Mj , for any
j = 1, . . . , R. System (1.4) always has a solution since the chain is absorbing. This solution
can be found by the method of inverse matrices.
The expected life time of party j is then given by the expression
( )

N
N
P j (1 − Pj )N −kj ×
k
Bj =
kj =Mj +1 kj j
(1.5)

R ∑
N
× P (X0 = k0 , X1 = k1 , . . . , XR = kR ) Bk0 , k1 , ..., kR .
ki =Mi +1
i ̸= j i = 0

The use of the introduced parameters in respective statistical data may allow practical
estimation of these parameters and give the answer to various interesting statistical questions.

2.2. Parameter estimation

7
A snapshot (census) may provide information data (X0 , X1 , . . . , XR ). Since this random
vector has a Multinomial distribution (1..2), the estimators for the probabilities Pj are
Xj
P̂j = , i = 0, 1, . . . , R . (1.6)
N
Additional information from the party archives will provide the average stay in each party
S̄j . Hence, the intensity of quitting party j equals
1
µ̂j = . (1.7)
S̄j

Further, from (1.1) we obtain that


λj Pj
= , i = 1, . . . , R .
µj P0
Combining these equalities, we find that the estimator of the intensities λj of the willingness
to join party j on behalf of a non-party person is

P̂j
λ̂j = µ̂j . (1.8)
P̂0
The estimators of parameters of the considered here party model are given by (1.6) - (1.8).

3. AN ELECTION MODEL
The party model of the previous section allows creating respective election model. When
using such models observers can predict results in elections. During election process people
vote for a party, coalition, or for individual. We call each of these separate units ”coalition”.
Coalition is a group of parties who agree to work in collaboration after the elections.
Assume the following configurations before the elections. There are registered C coalitions
for the peoples vote. A coalition may consist of one or several parties. The “Zero” party is
made by those voters who are not members of any party. They may be willing to vote with
probability p, or not vote at all with probability q = 1 − p. The vote of who decide to do so,
will go to coalition Cj with probability Qj proportional to the rates of sympathy towards
this coalition (see Remark 1), i.e.
λi1 + . . . + λiCj
Qj = , j = 1, . . . , C. (2.1)
λ1 + . . . + λR
Let Tj be the number of votes given in support of coalition Cj . With notation T0
for the number of individuals who did not voted, the distribution of the random vector

8
(T0 , T1 . . . , TC ) (the voting results) is multinomial with parameters Q0 = 1 − p, pQ1 , . . . , pQC .
Its probability generating function is given by the expression

T (z0 , z1 , . . . , zC ) = ((1 − p)z0 + pQ1 z1 + . . . + pQC zC )N , |zj | ≤ 1. (2.2)

The probabilities P (T0 = k0 , T1 = k1 , . . . , TC = kc ) equal to the coefficients in front of


the variables z0k0 z1k1 . . . zCkC in the power series expansion of the right hand side of (2.2). This
distribution allows calculation of various exact expressions for average characteristics related
to the specific electoral results.

3.1 Characteristics of the Election Model


Knowledge of the distribution of the random variables Tj (the number of votes given
in support of coalition Cj ) allows calculation of various interesting explicit and/or average
characteristics related to the specific electoral laws. For instance:
The average number of votes for coalition Tj equals N Qj , for j = 1, . . . , C, and
no voting people on average are equal to (1 − p)N.
The probability that coalition Tj does not survive the requirements to pass
a minimum percentage barrier γ % is given by the expression For instance, an important
characteristic is the probability that the coalitions can not survive the requirements to pass
the minimum percentage barrier γ (e.g. γ = 4% election barrier means P {Tj > .04 ×
#(all votes} is required for a coalition to be recognized as a political power until next
elections). For coalition Cj this probability equals
( )
Tj ∑N
Tj N k
P{ ≤ γ} = P{ ≤ γ|#(voters) = k} p (1 − p)N −k
T1 + . . . + TC k=0 T1 + . . . + TC k
( )

N
N k
= P {Tj ≤ γk} p (1 − p)N −k
k=0 k

∑ kγ ( )
N ∑
( )
k N k
= Qij (1 − Qj )k−i
p (1 − p)N −k .
k=0 i=0 i k
(2.3)
The complement to 1 of the probability given by (2.3) is the probability that this coalition
survives the elections. Since here the work is mostly with binomially distributed r.v.’s, all the
conventional approximations to the Binomial distribution (Poisson, normal) are legitimate
tools to simplify the calculations. We expect these details to grab the attention of skilled
applied probability experts.

9
In many cases, the votes from the losing coalitions are distributed proportionally between
the winning coalitions, according to the numbers of actual votes given for the winners. These
potential extra (donor) votes may increase the elected members from a wining coalition.
Respective conditional distributions are also available based on models discussed here. This
is a challenge for another study.
The probability that coalition Cj is the winner in the elections will be determined from
the requirement
Tj = max (T1 , . . . , TC ) . (2.4)

After we take into account equation (2.2) we easily find that the probability (2.4) to be
fulfilled equals


N
N!
βj = (1 − p)]k0 (pQ1 )k1 . . . (pQc )kC (2.5).
k0 !k1 ! . . . kc !
k j > k1 , . . . , k c
k0 + k1 + . . . + kC = N

It is intuitively clear that the coalition with the highest value of the probability Qj is the
expected winner. However, equation (2.5) allows to evaluate probabilities for any other
coalition to win.
One last remark here is that the results in votes are negatively correlated random vari-
ables. The correlation matrix among the coordinates of the random vector (T0 , T1 . . . , TC ) is
given by the entries
√ √
Qi Qj
ρTi ,Tj = −p for i ̸= j, i, j = 1, . . . , C.
1 − pQi 1 − pQj

Correlations with the number of no-voting eligible voters are determined by the expres-
sions v
u
u (1 − p)Qj
ρT0 ,Tj = −t for j = 1, . . . , C.
1 − pQj

4. A GOVERNING MODEL
The process of formation of governing coalition between the winners in the election has
a complex structure. Here we enter in the complexity of issues well described in chapters
3 and 4 by March (1994). It is a challenging task to make mathematical models based on
these descriptions of decision making.

10
We skip the process of formation of governing coalition between the winners in the elec-
tion. Usually, the coalition with major sits in the National Assembly (NA) takes responsibil-
ity to form a governing coalition. It negotiates the groups with smaller sits in the NA, and
as soon as gets the support of more than 50% of all the sits in the NA, the government is
ready to be formed. Let K be the critical number of sits supporting the governing coalition,
below which government is dissolved. This government survives as long as it is supported
by the majority of the members in the NA. Assume, the government is supported by G
coalitions, with X1 , . . . , XG representative sits in the NA. However, this is politics, and the
actual number of supporters may vary. The number of governmental supporters in each
group is presented by the random variables X1 , . . . , XG . In order to be functionally live, each
coalition needs a minimum number of supporters, i.e. to function a particular group j in the
NA it is needs to be fulfilled Xj ≥ Kj , j = 1, . . . , C. Here C is the number of coalitions
who took part in the elections, and Xj is the number of sits in the NA the j-th coalition
won in the elections. Some Xj ’s may be equal 0, if the coalition did not satisfy the election
requirements.
We assume that X1 +. . .+XG makes more than 50% of the sits in the NA, but the total of
any smaller subset from the random variables X1 , . . . , XG (the founders of the government)
makes less than 50% of the sits. Then the following reliability system with G subsystems
connected in series

→ [K1 − out - of − X1 ] ←→ . . . ←→ [KG − out - of − XG ] →

describes the functioning government Recent results on k − out − of − n reliability systems


which may suit studies of governmental unit as a political formation can be found in Cui
and Xie (2005).

4.1. Characteristics of the Governing Model


We model the government as a reliability system consisting of subsystems (the supporting
coalitions) connected in series, with variable number of functioning components in each
subsystem. If the NA works smoothly, without any attempts on behalf of the opposition
groups to call for a “non-confidence vote”, then the reliability of this government after t
units of time following its formation is given by the rule
Xj ( )

G ∑
Xj
R(t) = (1 − Fj (t))i (Fj (t))Xj −i . (3.1)
j=1 i=Kj i

11
Here by Fj (t) we denote the distribution function of the random duration time a member
of the NA from coalition j keeps his/her loyalty to the governmental formation. Decisions
are assumed independent between the members of the NA.
But this is politics, and the life is dynamically changed. There are non-confidentiality
votes against the government. In case that the opposition parties call for non-confidentiality
vote for this particular government, the results from the vote may be modeled by making use
of the models from sections 2 and 3. Such votes are binary (Pro or Con), and each coalition
may have its probability pj for members to vote “Pro”. The random number Nj of “pro”-
votes in each coalition Cj is binomial of parameters (Xj , pj ), where Xj is the number of sits
the j th coalition keeps in the NA. Therefore, the total probability rule can be consecutively
applied. In this way we obtain that the chance to survive a non-confidentiality vote at time
t has probability
Xj ( )

C ∑ ∑ Xj
P (N1 + . . . + NC ≥ K; t) = (1 − Fj (t))i (Fj (t))Xj −i ×
j=1 k1 +...+kc ≥K i=Kj i
( )
(3.2)
i k
× pj j (1 − pj )i−kj .
kj

If the calls from opposition for non-confidentiality vote to the government form a Poisson
process of certain parameter γ > 0, then for any t ∈ (0, T ) the reliability function of the
coalition is solution of the recursive equation
∫ t
P (N1 + . . . + NC ≥ K; t) = R(t) + γP (N1 + . . . + NC ≥ K; t − τ ) e−γτ dτ , (3.3)
0

where R(t) is determined by (3.1).


If the elections are scheduled periodically after expiration of time T , the average life
time of the government will be presented by the quantity
∫ T
µT = E(L|T ) = P (L > t) dt . (3.4)
0

The probability that the government will complete its mission until the next scheduled
elections (time of duration T ) in presence of non-confidentiality activities of intensity γ
satisfies the inequality
P (L ≥ T ) ≥ R(T )e−γT .

It is intuitively clear that the more support in the NA to the government, the more
stable this government is. However, its stability may depend on strong leadership, on external

12
factors. Drift in economics, in state financial status, strict fulfilment of pre-election promises,
and numerous other factors may make the governmental modeling attractive target to applied
probability researchers and statisticians. This is a specific risk theory process working in finite
time. Each government comes with certain amount of credits. When its credits expire or
wasted, it should go. The ruin theory may suit in such considerations. The uncertainty in
government overall motion is possibly suitable to be described as a diffusion process. Such an
approaches are promising, and may provide ways to consideration additional characteristics
of the government as a unit. It is worthy to be worked out.

5. CONCLUSIONS
Political processes offer challenging area of applications of various mathematical modeling
approaches and theories. Usually they are modeled by simulation of particular situations, or
by models of games theory. We propose probability models for traditional, democratic polit-
ical environment, which stick close to the processes of formation of political activities and to
their natural components. Inclusion of the uncertain political elements into specific proba-
bilistic model seems simple. The obtained in this approach analytical results are promising.
They offer a field for lot of discussions, continuing statistical considerations various kinds of
applications and further ideas of development.
There are other existing political models (monarchy, totalitarian, military regimes, etc.)
than the considered democratic model. We are confident that a detailed process analysis can
make these processes available for successful respective quantitative modeling.

BIBLIOGRAPHY

Cioffi-Revilla, C. (2009) Simplicity and reality in computational modeling of politics, Com-


put. Math. Org. Theory, Springer, 15 (1), 26-46.

Clark, T. D., Larson, J. M., Mordeson, J. N., Potter, J. D., and Wierman M. J. (2008)
Applying Fuzzy Mathematics to Formal Models in Comparative Politics, Springer.

Cui, L., and Xie, M. (2005) On a generalized k-out-of-n system and its reliability, Int. J.
Sys. Sci. 36 (5), 267-274.

Doyle, D. R. (2000) Political Complexity: Nonlinear Models of Politics, University of Michi-


gan Press.

Esa, S., and Dimitrov, B. (2009) Reliability Modeling of Political Processes, Proc. VI Inter-

13
nat. Conf. MMR’2009, Moscow, pp. 268-272.

Esa, S., and Dimitrov, B. (2011) Survival Models of Some Political Processes, Proc. VII
Internat. Conf. MMR’2011. Edts L. Cui & X. Zhao. Beijing Inst. of Technology Press,.
721-726.

Gross, D., and Harris, C. M. (1998). Fundamentals of Queueing Theory, New York: Wiley

Hafer, C. (2007) Contests over Political Authority, New York University Mimeo .

J. G. March (1994) A Primer on Decision Making: How Decisions Happen. New York, NY:
The Free Press.

Monroe, K. R. (1997) Contemporary Empirical Political Theory, Princeton University


Press, 1997.

Morrow, J. D. (1994) Game Theory for Political Scientists, University of California Press.

Morton, R. B. (1999) Methods and Models: A Guide to the Empirical Analysis of Formal
Models in Political Science, Cambridge University Press, 1999.

Obretenov, A., and Dimitrov, B. (1979) Handbook in Queuing, Sofia: Nauka I Izkustvo.

Ordeshook , P. C. (1986) Game theory and political theory: an introduction, Cambridge


University Press.

Taylor, A. D. (1995) Mathematics and politics: strategy, voting, power and proof, Springer.

14

You might also like