ISSN 1392-8619 print/ISSN 1822-3613 online

http://www.tede.vgtu.lt 531
EXPERIENCE OF THE GAME THEORY APPLICATION IN
CONSTRUCTION MANAGEMENT
Friedel Peldschus
Vilnius Gediminas Technical University, Saulėtekio al. 11, LT-10223 Vilnius, Lithuania
E-mail: friedel@peldschus.net; friedel.peldschus@st.vgtu.lt
Received 12 August 2008; accepted 25 November 2008
Abstract. Te game theory allows mathematical solutions of confict situations. Besides the fairly
established application to economical problems, approaches to problems in construction operation
have been worked out. An overview of applications is given. Solution strategies for such engineering
problems are collected. Furthermore, concrete application examples are presented and an overview
of further potential applications is given. Solutions of two-person zero-sum games are discussed as
well as approaches to fuzzy games.
Keywords: construction management, selection of variants, game theory, two-person zero-sum
games, fuzzy games.
Reference to this should be made as follows: Peldschus, F. 2008. Experience of the game theory
application in construction management, Technological and Economic Development of Economy
14(4): 531–545.
1. Introduction
Complex decisions based on incomplete information about infuence factors, relations and
possible events are needed in construction operation. Engineering work is aiming at planning
in advance the necessary actions for solving such problems. Machines, tools, processes are
chosen and concepts for their coordination are made.
Looking in detail at typical decisions in construction operation it can be observed that
this incompleteness of the information is ofen neglected. Either the most favourable param-
eters are used and a over-dimension or other economical disadvantages are seen as given, or
simplifed deterministic models are applied, possibly modifed according to experiences but
not representing the situation precisely enough. Tis approach, in general, leads to a feasible
solution but not necessarily to the most favourable one.
TEchNOLOGIcAL ANd EcONOmIc dEVELOpmENT OF EcONOmY
Baltic Journal on Sustainability
2008
14(4): 531–545
doi: 10.3846/1392-8619.2008.14.531-545
532 F. Peldschus. Experience of the game theory application in construction management
Te investigation of principles is aiming at the elimination of information incompleteness.
Te application of principles allows to assess decisions with respect to the necessary actions
and to give an orientation for selecting them.
A simple assessment for all possible results of decisions is not always sufcient, because
for every decision diferent, mutually exclusive results can be obtained. As it is not known
beforehand, which result is actually going to occur, decision criteria considering the totality
of all possible results are needed.
As a solution to this problem, the application of the Game Teory is proposed. Te Game
Teory deals with the representation of confict situations and their resolution. Tus, it allows
choosing the most favourable one out of a set of diferent behaviours. Application examples for
problems of construction operation, that have been solved by this theory, are presented.
2. Teory of the matrix games
For solving engineering problems matrix games (Peldschus et al. 1983a; Peldschus 1986)
are used. Tese games are two-person zero-sum games (Peldschus and Zavadskas 1997;
Zavadskas et al. 2004; Hollert 2006).
Γ
( )
S S A
1 2
, , , (1)
with
S
i 1
for i = 1, ..., m strategies of player I,
S
j 2
for j = 1, ..., n strategies of player II,
and the pay-of function for players I and II:
S S S
S a a a
S a a a
S a a a
n
n
n
m m m mn
21 22 2
11 11 12 1
12 21 22 2
1 1 2



 

For these games ideally a saddle-point solution (simple min-max principle) or a strategy
combination (extended min-max principle) is obtained.
Simple min–max principle
α max min
i j
ij
a ; β min max
j i
ij
a . (2)
If α β γ , a saddle point with pure strategies (one optimal strategy for each player) is
obtained as solution – trivial solution.
Extended min–max principle
An equilibrium point with mixed strategies is calculated (combination of strategies) as:

max min ( , ) min max , ,
* *
i j i j
A s s A s s A s s
1 2 1 2 1 2

( )

( )
ν
. (3)
533 Technological and Economic Development of Economy, 2008, 14(4): 531–545
Further specifc solution methods for games against nature were discussed in detail in
Zavadskas et al. (2002, 2003, 2004), Zavadskas and Sivilevičius (2005).
2.1. Applications
2.1.1. Selection of variants
Te main application of matrix games is the selection of variants, which is a problem of
multi-criteria decisions (Peldschus et al. 1983b, 2005, 2007a; Antuchevičienė et al. 2006;
Meszek, 2001, 2004, 2007; Ustinovichius et al., 2007; Zagorskas and Turskis 2006; Zavadskas
and Kaklauskas 2007; Zavadskas and Turskis 2008).
For describing the problem the variants are assigned to the strategies of player I and
the criteria to the strategies of player II. For the pay-of function dimensionless evaluation
numbers are used in simple cases. Such numbers describe the situation only coarsely. It is
therefore sensible to use real characteristic values. As such values have diferent dimensions,
their efectiveness is not comparable. In order to compare the characteristic values they are
mapped on the interval [1, 0] or [1, ~0]. Depending on the kind of problem, there are sev-
eral options for transforming the characteristic values. Generally, a distinction can be made
between linear and non-linear transformations (Peldschus 2007a; Brauers and Zavadskas
2007; Ginevičius 2008; Migilinskas 2003; Zavadskas et al. 1994; Zavadskas and Kaklauskas
2007; Zavadskas and Turskis 2008).
Variants with one variable
For the selection of variants, m variants and n criteria can be used. Te variants are grouped
for solving of the problem. Labelling the variable with X the following variants are given:
VAR.1 = VAR (X
1
), VAR.2 = VAR (X
2
), ... , VAR.M = VAR (X
m
). (4)
When generalised, this can be written as:
VAR.i = VAR (X
i
). (5)
For these variants the results
a a X
ij i i

( ) (6)
with respect to the criteria K K K
j n

¦ ¦
1
, ,  are calculated and added to the decision
matrix:
VAR.1 = VAR (X
1
)
K K K
a a a
a a a
a a a
n
n
n
m m mn
1 2
11 12 1
21 22 2
1 2



   


VAR.2 = VAR (X
2
)
VAR.M = VAR (X
m
)

534 F. Peldschus. Experience of the game theory application in construction management
With the variants VAR.i = VAR (X
i
) as strategies of player I, the criteria K K K
j n

¦ ¦
1
, , 
as strategies of player II and the transformed values a a X
ij i i

( ) as pay-of function, a solu-
tion as two-person zero-sum game is possible.
Example: Multiple-criteria analysis in the construction of motorways
Te steady growth of road trafc in Germany requires the extension of roads and motorways
network. In addition to the construction of new routes, the upgrading of the existing network
is of a particular importance. Tis includes the refurbishment of damaged motorway seg-
ments and the expansion from 4 to 6 lanes.
Te following variants for the 6-lane expansion of a motorway in Turingia were ana-
lysed:
Variant 1 – Reconstruction with a concrete surface, modifed axis position
and gradient;
Variant 2 – Reconstruction with an asphalt surface, modifed axis position
and gradient;
Variant 3 – Reconstruction with a concrete surface, modifed axis position
and identical gradient;
Variant 4 – Reconstruction with an asphalt surface, modifed axis position
and identical gradient;
Variant 5 – Reconstruction with a concrete surface, identical axis position
and identical gradient;
Variant 6 – Reconstruction with an asphalt surface, identical axis position
and identical gradient.
Te selection was made according to the criteria: construction costs, construction period,
durability, environmental protection and economic efciency. It was aimed at fnding the
optimal solution that fulfls the mentioned criteria best.
Te values for the criteria costs, construction period, length of the route, noise protection
originate from operational documents. For the durability, statistical data were collected in a
diploma thesis. Te following values were used:
Table 1. Multiple-criteria analysis in the construction of motorways
Variant
Costs
[mln €]
(min)
Durability
[years]
(max)
Noise level
[db(A)]
(min)
Length of route
[m]
(min)
Construction period
[days]
(min)
1 12.49 30 62.61 1088 761
2 12.37 20 59.61 1088 746
3 11.10 27 62.62 992 669
4 10.98 18 59.62 992 654
5 11.02 24 62.83 998 700
6 10.90 16 59.83 998 685
535 Technological and Economic Development of Economy, 2008, 14(4): 531–545
Te numerical solution was performed by the sofware LEVI which has been developed
in a co-operation with HTWK Leipzig and VGTU Vilnius. Te values of Table 1 are mapped
onto a dimensionless interval using the non-linear transformation (Peldschus):
b
a
a
ij
ij
i
ij

j
(
,
,
,
\
,
(
(
(
min
3
if min
i
ij
a is advantageous,
(7)
b
a
a
ij
ij
ij
i

j
(
,
,
,
\
,
(
(
(
max
2
if max
i
ij
a is advantageous.
Te solution is made according to the extended min-max principle formula (3).
Te obtained solution vectors are:
S

= ( 0.254; 0 ; 0.746 ; 0 ; 0 ; 0 ) and S

= ( 0 ; 0.612 ; 0 ; 0 ; 0.388 ).
Variant 3 gets the highest rating and can therefore be declared as the optimal variant. A
comparative calculation with the Laplace criterion, treating all criteria as equally important,
confrms this result. Te game-theory equilibrium is formed by the criteria durability and
construction period. Other criteria are of minor importance (Peldschus 2005).
Te calculation of the optimal number of partial assembly lines for work-cycle produc-
tion (Peldschus 1975), the selection of wall structures in residential construction (Peldschus
et al., 1983b), the estimation of retail centres infuence on the city structure (Zagorskas and
Turskis 2006), the selection of buildings’ maintenance contractor (Zavadskas et al. 2005) are
examples of game theory applying in construction.
Variants with several variables
If the problem cannot be addressed to by a formation of variants with one variable, there
is the possibility to include several variables in the formation of variants. m variants and n
criteria can be considered. Te formation of variants is performed for all combinations of the
considered variables. With the variants being X X
k 1
, ,  and the possible values for X
l
being
v z
l l

¦ ¦
1,..., , the following variants are obtained: VAR.1 = VAR (X X X
k
1
1
1
2
1
, , ,  ), VAR.2 =
VAR ( X X X
k
2
1
1
2
1
, , ,  ), ..., VAR.M = (X X X
Z Z Zk
k
1
1
2
2
, , ,  ). Te variant i is derived from
i V Z Z V Z V
k k k k

( )
+ −
( )
+ + −
( )
+
− − −
  1 1 1
1 1 2 2 1 1
ν . (8)
In total, the following variants are obtained

M


Z
i
i
k
1
. (9)
For  these  variants  the  results  a
ij
  for  the  criteria  K K K
j n

¦ ¦
1
, ,    are  calculated  and 
added to the decision matrix. Similarly to the formation of variants with one variable, this 
decision matrix is transformed into a two-person zero-sum game by considering the variants 
VAR.i as strategies of player I and the criteria K
j
as strategies of player II. Te values of the
pay-of function are obtained by transformation to dimensionless values.
536 F. Peldschus. Experience of the game theory application in construction management
Consideration of the frequency of the variants
If the calculation is to be made for a serial production with diferent constructions, the matrices
for each construction can be multiplied with their frequency. With t e ∈
¦ ¦
1,..., constructions,
the matrices A
t
are obtained. With the frequency g
t
this gives:
′ × A g A
t t t
. (10)
Adding all matrices
′ A
t
and dividing by the number of constructions, the weighted matrix
A
*
is obtained:

A
A
g
t
t
e
t
t
e
*



1
1
. (11)
Afer transformation to dimensionless values this weighted matrix A
*
can be used as
pay-of function for a two-person zero-sum game. A production variant, taking into account
the frequency of the diferent constructions, is obtained as a result. Tis method has been
applied to dimension of the technological line for producing pre-fabricated concrete parts
(Altmann and Peldschus 1987).
2.1.2. Calculation of correction values
Providing informative numbers for planning is an important problem in construction prepa-
ration. Tese values (parameters) are determined by analyses. Such analyses are, however,
associated with uncertainties, which infuence planning and lead to diferences between
planned and actual eforts. Tey can infuence the capacity assignment, material consump-
tion planning and structural parameters. When the performance is to be increased due to
deviations from the planning, then the capacity of the dimensioned system is exceeded. A
lack of labour or material can be given as an example. When the output is to be decreased,
the intended productivity cannot be achieved.
An increase of the planning eforts can only partly compensate these uncertainties. Te
problem is very obvious in the reconstruction of buildings. In general, problems for which dif-
ferences between planned and actual eforts are found should be interpreted in this context.
In order to compensate for diferences between planned and actual eforts, reserves have
been foreseen by construction companies. Tese reserves are based on estimations, which
can, however, be not satisfactory. Terefore attempts were made to determine this variable
mathematically. As a result, the amount of the reserve is considered as a planning strategy.
Te uncertainties of the analysis of the building conditions are forming the opposing side. A
pay-of function has to be determined for both according to the game theory. If the amount
of the reserve equals the performed construction eforts, the target is reached. In this case the
value of the pay-of function is set to 1. In all other cases there is a diminution by a quotient
smaller than 1.
537 Technological and Economic Development of Economy, 2008, 14(4): 531–545
a
P
Z
P Z
ij
i
j
i j

+
+

1
1
for ; a
Z
P
P Z
ij
j
i
i j

+
+
1
1
for  ; (12)
and i = 1, ..., n; j = 1, ... m;
P
i
– planning reserve; Z
j
– diference;
m – number of the analysed strategies (strategies of player I);
n – number of diference classes (strategies of player II).
Tis, the two-person zero-sum game is defned. It is solved according to the Bayes cri-
terion (Peldschus 1986).

S S S q a q
i i
i
j ij
j
n
j
j
n
1 1 1
1 1
1
*
/ max ∩
j
(
,
\
,
(

¦
¦
¦
¦
¦
¦
¦
¦
¦
¦

∑ ∑
. (13)
Example: Plumbing in reconstruction of buildings in Leipzig
For the numerical calculation, the required range is defned. As indicated in Table 2, in this
case the range was set from a shortfall of – 20% to an exceedance of 50% based on statistical
investigations.
Table 2. Plumbing in reconstruction of buildings in Leipzig
Z
j
P
i
–0.2 –0.1 0.0 0.1 0.2 0.3 0.4 0.5
0.5 0.064 0.072 0.053 0.059 0.064 0.173 0.187 0.120
0.4 0.068 0.077 0.057 0.063 0.067 0.186 0.200 0.112
0.3 0.074 0.083 0.062 0.068 0.074 0.200 0.171 0.104
0.2 0.080 0.090 0.067 0.073 0.080 0.185 0.186 0.096
0.1 0.087 0.098 0.073 0.080 0.073 0.169 0.157 0.088
0.0 0.096 0.108 0.080 0.073 0.067 0.154 0.143 0.080
–0.1 0.106 0.120 0.720 0.065 0.060 0.138 0.128 0.072
–0.2 0.120 0.106 0.064 0.058 0.053 0.123 0.114 0.064
Te calculated planning reserve is 20%. Such correction values may be also calculated for
other crafs (Höher and Peldschus 1981).
2.1.3. Complex evaluation
Te complex assessment is feasible with a game-theory model, if single evaluations for the
defned criteria are available and if the evaluation of the interrelation cannot be performed
538 F. Peldschus. Experience of the game theory application in construction management
by summation or by multiplication due to lacking validity.
A weighting of the analysed variants in terms of a global goal is obtained with the game-
theory model. Tis result may be interpreted as a ranking and it is valid in the sense of shares.
Te game value may be understood as grade of efectiveness with respect to the global goal
and may be assigned to a defned classifcation.
Te quality of the results depends on the validity of the pay-of function. In the simplest
case it can be defned by evaluation numbers (point system) or in case of a more detailed
input by efective values.
Example: Multi-criteria decision support in facilities management
Nine comparable buildings were chosen for this survey, outliers have not been considered.
Tese buildings are characterized according to water consumption [m³/m²], heat require-
ment [kWh/m²] and electricity consumption [kWh/m²]. Heat requirement and electricity
consumption have got the same unit of measurement. But their values cannot be summed up,
because prices for heat requirement and electricity consumption are not the same. Tus, the
addition would distort the statement. Te consumption costs also are not relevant because
of the partly considerable price diference in various regions.
As non-monetary criteria, exemplary elements of use were considered which express
customer and employee friendliness.
Table 3. Multi-criteria decision support in facilities management
Location
Air
condi-
tioning
Automat-
ic doors
Eleva-
tor
Consumption
of electricity
[kWh/m
2
]
Kind
of heat-
ing
Heat con-
sumption
[kWh/m
2
]
m
3
wa-
ter/m
2
Building
1
Yes Yes No 126.1
District
heating
39.76 0.08
Building
2
No No No 104.4
District
heating
44.08 0.12
Building
3
No No No 161.2 Gas 115.40 0.09
Building
4
No Yes No 130.0 Gas 101.00 0.13
Building
5
No No No 50.0 Gas 153.14 0.07
Building
6
No No No 93.7 Gas 70.34 0.06
Building
7
Yes Yes No 125.5 Gas 68.64 0.09
Building
8
Yes Yes No 155.9 Gas 119.16 0.12
Building
9
Yes Yes No 133.9 Gas 33.31 0.18
Te numerical processing is done with the help of the computer program LEVI 3.0
(Peldschus et al. 2002; Zavadskas et al. 2002, 2004; Zavadskas and Sivilevičius 2005; Zavad-
539 Technological and Economic Development of Economy, 2008, 14(4): 531–545
skas and Turskis 2008) jointly developed by VGTU Vilnius and HTWK Leipzig. Tere, air
conditioning/ventilation was taken into account with 50%, automatic doors with 10%, and
elevator with 40% respectively.
Because of the diferent dimensions or statements, the values in the matrix cannot be
compared. Terefore a transformation is performed to an interval with dimensionless values,
the interval [1; ~0]. Te computer program LEVI 3.0 allows the following transformations:
normalization of vectors, linear transformation, relative variance and non-linear transforma-
tion. For the task in question the non-linear transformation (formula 7) was chosen. Tis
decision is justifed because of the values range of the criteria examined. With values for the
range ≤ 2 min the other transformations are only applicable to a certain extent or not at all.
Te game-theoretical equilibrium point (formula 3) of the game is formed from the
evaluation of the buildings 1, 5 and 9. Of these, building 1 yields the best valuation. However,
the diference to building 5 is insignifcant. Terefore both buildings can be considered as
approximately equal. With regard to criteria, the equilibrium point arises from the efect of
the criteria water consumption, heat consumption and electricity consumption. Because of
their functional course, equipment and furnishing are not included in the solution.
Te calculated value of the game is υ= 0.434, and can be rated as quality level of the
group of buildings. Te value can be used for comparison between several groups of build-
ings (Peldschus and Reichelt 2002).
Tis method has also been applied to assessing the obstructions of construction works in
the reconstructed buildings of the chemical industry (Simon 1982), the assessment of noise
control measures (Peldschus 1986) and the sensitivity analysis in land evaluation (Meszek
2004).
3. Fuzzy games
3.1. Teory of fuzzy games
New approaches can be developed by applying the theory of fuzzy sets to the game theory.
Tereby it is feasible to develop numerical solutions to problems which cannot be defned
exactly.
A fuzzy set A in X is a set of ordered pairs.

A x x x X
A

( ) ( )

¦ ¦
, / µ
. (14)
Hereby µ
A
x
( ) is the degree of association of X with the fuzzy set A. Te entirety of the
values µ
A
x
( ) for all potential elements x of X constitutes the association function. Tis as-
sociation function µ
A
x
( ) can reach all values of non-negative real numbers. Generally it is
oriented to the interval [0; 1], 0 1 ≤
( )
≤ µ
A
x .
The association function  µ
A
x
( ) can be interpreted as:
– degree of the contentedness of x in A;
– degree of association of x with A;
– degree of the possibility of incidence of x;
– degree of compatibility of x with a given characteristic;
– degree of the engagements of x for a target;
540 F. Peldschus. Experience of the game theory application in construction management
– degree of comprehension of x in an issue;
– degree of concordance of x with a given statement.
Te interpretation of the association function is task dependent. Hence it has to be re-de-
fned for every problem. For a fuzzy decision association function µ
A
x
( ) indicates to which
grade each element x complies with the given requirements.
An element x X ∈ represents an optimal fuzzy decision, if x has the maximum degree
of association to A.
Diferent concepts for determining the association function are known. Piecewise linear
functions are ofen used. Functions of an S-shape are more suitable for solving engineering
problems. Te interpolating cubic spline function has proven to be suitable for practical
problems of multi-criteria decisions (Omran 1988).
With the nodes (x
0
; 0), (x
D
; 0.5), (x
M
; 1) and the boundary conditions ′
( )

( ) µ µ
A A M
x x
0
0
and
′′
( )
µ
A D
x 0 two third-grade polynomial are obtained. Tese are attached continuously
in x
D
in a way that they can be diferentiated twice. Tus for x x
m
>
0
the following approach
is obtained:
µ
A
D
D
x
x x
Ax Bx Cx D x x x
Ex Fx Gx H x x x
( )

+ + + ≤ ≤
+ + + ≤ ≤
0
0
3 2
0
3 2
for
for
for
;
;
MM
M
x x
;
. 1 for ≤
¦
¦
¦
¦
¦
¦
¦
(15)
Te coefcients (A,... , H) are calculated from an equation system (14), which is defned
by the requirements to
µ
A
x
( ) of continuity, existence of the frst and second derivatives, and
the selection of the nodes.
G x x x
1 0
3
0
2
0
0
( )
+ + + A B C D ;
G x x x
D D D 2
3 2
0 5
( )
+ + + A B C D . ;

G x x x
D D D 3
3 2
0 5
( )
+ + + E F G H .
;

G x x x
M M M 4
3 2
1
( )
+ + + E F G H
;
G x x
5 0
2
0
0
( )
+ + 3A 2B C ; (16)

G x x
M M 6
2
2 0
( )
+ + 3E F G ;


G x x x x
D D D D 7
2 2
2 2 0
( )
+ + − − − 3A B C 3E F G ;


G x x
D D 8
2 2 0
( )
+ − − 6A B 6E F .

Te equation system (G
1
), ..., (G
8
) is not singular and possesses a unique solution. In order
that µ
A
x
( )
⊂[0; 1] and it is also monotone in x
i
, the following condition must be fulflled,
which is usually the case:

− + ≤


≤ + 1 2 1 2
0
x x
x x
M D
D
. (17)
Complications only arise, if x
D
is near a
0
or x
M
.
541 Technological and Economic Development of Economy, 2008, 14(4): 531–545
For the combination of the theory of matrix games and the theory of fuzzy sets a three-
step model was proposed (Omran 1988). Te fuzzy game Γ
µ
, which contains partly or wholly
fuzzy information, can be described as follows:
Γ
µ
µ µ µ
¦ ¦
( , ); ( , ); ( , ) S S a
i i j ij ij ij 1 1 2

, (18)
with S
i 1
for i = 1, ..., m strategies of player I ;
µ
1i
for i = 1, ..., n association function for the strategies of player I;
S
j 2
for j = 1, ..., n strategies of player II;
µ
ij
for j = 1, ..., n association function for the strategies of player II
with respect to strategies of player I ;
a
ij
for i = 1, ..., m pay-of function j = 1, ..., n;
 µ
ij
for i = 1, ..., m association function for the pay-of function j = 1, ..., n.
Te transition from game Γ to game
Γ
µ
is done in 3 steps:
1
st
step
On the set of strategies of player I a fuzzy set is defned. Te set of the strategies and the
set of the criteria, which quantitatively describe the set of the strategies, are known. For each
criterion the association function (15) is calculated, i.e. the norm values are transferred to
association values. Tus, for each strategy of player I an association value for several criteria
is obtained. As compensation between the criteria is admitted, a summation is done by cal-
culating the arithmetic mean (Laplace-criterion).

µ µ
1
1
1
i il
i
L
L


. (19)
2
nd
step
In the second step the strategies of player II are considered. Also, on the set of strategies
of player II fuzzy sets are defned and association values are calculated according to (15).
Te mapping of sets is done in matrix format. Tis frst gives a basic matrix for the game Γ
µ
to be solved. In contrast to the frst step, this matrix is interpreted with respect to the game
theory.
3
rd
step
The third step is combination of step 1 and 2. Thus, it is an intersection of the strategies 
of players I and II.
A widely accepted approach is based on Bellmann and Zadeh (1970), in which the fuzzy
decision is defned as the intersection of the fuzzy sets for the fuzzy goals Z and the fuzzy
set of the restrictions R. For the average of the fuzzy sets Z and R the association function is
defned point wise by the operator

µ µ µ
ij Z R
x x x
'
( ) min ( ), ( )
[ ]
. (20)
As a result, the actual fuzzy game matrix is obtained. For the solution, the min-max-
principle, adopted from the classic game theory, is used.
542 F. Peldschus. Experience of the game theory application in construction management
3.2. Applications
Te developed algorithm for fuzzy matrix games is a fuzzy concept for multi-criteria deci-
sions. Tis concept was developed for considering internal and external infuences.
Internal infuences have the character of experiences and they apply until the system
comes in use, i.e. in the construction or production phase. External infuences describe a
new quality. Tey have the character of prognosis and represent the utilisation phase. Tus
an algorithm is available to aggregate also quality characteristic with a hierarchical structure,
hereby admitting diferent phases.
For practical applications the construction variants become the strategies of player I.
All real construction variants, which could be discussed for the problem to be solved, are
considered. Tese construction variants are evaluated with respect to several parameters.
Te parameters represent the internal infuences and relating to the construction phase. As
specifc parameters, the following examples can be given: the required space, the construction
time, the construction costs, the required capacity etc. For these criteria values, association
values are calculated according to (13). Compensation between the parameters is admitted,
so that the application of a compensatory parameter (17) is appropriate. With the calculation
of the association values for the strategies of player I, their efective share is expressed. Te
strategies of player II contain the infuences from the use of the building. Tese include: the
resistance against failure of the construction, the amortisation, the operational and energy
costs. Tey represent the external infuences. Te interrelation of the strategies of player II
with the strategies of player I are described with the minimum operator (18). A matrix, which
describes the fuzzy game, is obtained as a result. Te solution is found using the min-max
principle (3). It has a strategical character.
Specifc investigations have been performed for variant evaluation of cable routes in the
chemical industry and for water supply systems (Peldschus and Zavadskas 2005) as well as
for multiple-criteria decisions in risk management (Reichelt and Peldschus 2005).
4. Results
Te game theory applications show new possibilities of the solution of confict situations
in construction, which are constituted by the consideration of several, partly contradictory,
objectives. Tese serve the attempt for gaining maximum beneft in conducting construction
works. It is known which difculties are associated with such a task, even in case of a number
of comparable situations. Terefore approaches, which are achieved by the application of
mathematical methods, are shown. Tese methods difer substantially from those applied so
far. Te result of these investigations is the selection of the optimal variant in consideration of
several criteria and the compliance with practical conditions associated uncertainty, lack of
information and fuzziness. With fuzzy games a lot of new solution possibilities for problems
from technical felds, economical felds, investment planning, etc. arise. Te solution of fuzzy
games forms a new quality of decisions, which represent a high degree of complexity.
543 Technological and Economic Development of Economy, 2008, 14(4): 531–545
Te application of the theory of matrix games may seem simple. It was, however, proven,
that every real confict situation is difcult to be analysed due to many, also complex, infu-
ences. Every formal description of a matrix game exhibits particularities, which need to be
considered in the solution as well as in the formalisation of the pay-of function. Te purpose
of the results is always the decisive element. Here a clear and unambiguous defnition of the
problem is necessary. Tis issue becomes more and more important in the light of increasing
complexity of construction preparation, and the need for objectiveness gains new dimen-
sions. Te application of the theory of matrix games will contribute to the improvement of
the objectiveness in construction preparation.
References
Altmann, P.; Peldschus, F. 1987. Rechnergestützte Optimierung bei der technologischen Bemessung von
Fertigteillinien für die Betonelementeherstellung, in 11. Internationaler Kongress über die Anwend-
ung der Mathematik in den Ingenieuwissenschafen, Weimar, Wiss Zeitschrif der Hochschule für
Architektur und Bauwesen Weimar, Reihe B Hef 4, 231–233.
Antuchevičienė, J.; Turskis, Z.; Zavadskas, E. K. 2006. Modelling renewal of construction objects applying
methods of the game theory, Technological and Economic Development of Economy 12(4): 263–268.
Bellman, R. E. and Zadeh, L. A. 1970. Decision-making in a fuzzy environment, Management Science
17(4):141–164.
Brauers, W. K.; Zavadskas, E. K. 2007. Normalisation in decision-making methods, Editorial. International
Journal of Management and Decision Making 8(5-6): 441–444.
Ginevičius, R. 2008. Normalization of quantities of various dimensions, Journal of Business, Economics
and Management 9(1): 79–86.
Höher, G.; Peldschus, F. 1981. Eine Methode zur Verbesserung von Rekonstruktionsvorhaben im Woh-
nungsbau, in 9. Internationaler Kongress über die Anwendung der Mathematik in den Ingenieuwis-
senschafen, Weimar, Berichte Hef 1, 46–47.
Hollert, M. J. 2006. Einführung in die Spieltheorie. Springer Verlag Berlin. 429 S.
Meszek, W. 2001. Estimation of land plots using game theory, Technological and Economic Development
of Economy 7(2): 62–68.
Meszek, W. 2004. Usage theory of the plays for the plot value inquiry, Technological and Economic De-
velopment of Economy 10(1): 40–46.
Meszek, W. 2007. Uncertainty phenomenon on property valuation, International Journal of Management
and Decision Making 8(5–6): 575–585.
Migilinskas, D. 2003. Te infuence of normalization methods selection in construction including game
theory adaptation, Technological and Economic Development of Economy 9(2): 73–79.
Omran, J. 1988. Entscheidungsunterstützende Methoden in der bautechnologischen Vorbereitung unter
Berücksichtigung von unscharfen Spielen – Fuzzy Games. Dissertation A, Technische Hochschule
Leipzig. 137 p.
Peldschus, F. 1975. Die optimale Anzahl der Teiltaktstraßen für die Fließfertigung, Wissenschafliche
Zeitschrif der Hochschule für Bauwesen Leipzig 1/75: 53–55.
Peldschus, F. 1986. Zur Anwendung der Teorie der Spiele für Aufgaben der Bautechnologie. Dissertation B,
Technische Hochschule Leipzig. 119 p.
Peldschus, F. 2005. Multiple-Criteria Analysis in the Construction of Motorways, Technological and
Economic Development of Economy 11(1): 32–35.
544 F. Peldschus. Experience of the game theory application in construction management
Peldschus, F. 2007a. Te efectiveness of assessments in multiple criteria decisions, International Journal
of Management and Decision Making 8(5–6): 519–526.
Peldschus, F. 2007b. Games-theory solutions in construction operation, In Modern Building Materials,
Structures and Techniques, in M. J. Skibniewski, P. Vainiunas, E. K. Zavadskas (eds.). Te 9th Inter-
national Conference, May 16–18, 2007. Vilnius, Lithuania, Selected Papers, 348–353.
Peldschus, F.; Messing, D.; Zavadskas, E. K.; Ustinovičius, L.; Turskis, Z. 2002. LEVI 3.0 – Multiple Cri-
teria Evaluation Program under uncertainty, Technological and Economic Development of Economy
8(1): 3–12.
Peldschus, F.; Reichelt, B. 2002. Multicriteria Decision Support in Facility Management, in Proceedings
Volume of the Meeting of the European Working Group of Multiple Criteria Decision Aiding, Coimbra
PT, editor European Commission Joint Research Centre.
Peldschus, F.; Vaigauskas, E.; Zavadskas, E. K. 1983. Technologische Entscheidungen bei der Berücksich-
tigung mehrerer Ziehle, Bauplanung–Bautechnik 37(4): 173–175.
Peldschus, F.; Zavadskas, E. K. 1997. Matriciniai lošimai statybos technologijoje ir vadyboje [Matrix Games
in Building Technology and Management]. Vilnius: Technika. 134 p. ISBN 9986-05-332-3.
Peldschus, F.; Zavadskas, E. K. 2005. Fuzzy Matrix Games Multi-Criteria Model for Decision-Making in
Engineering, Informatica 16(1): 107–120.
Reichelt, B.; Peldschus, F. 2005. Te Application of Multi-Criteria Decision Analysis (MCDA) in Risk
Management of Civil and Environmental Engineering Projects, Foundations of Civil and Environ-
mental Engineering, Poznan University of Technology, 6: 159–173.
Simon, L. 1982. Eine Methode zur Erfassung, Bewertung und Klassifzierung von Einfüssen, die den Bauab-
lauf der Rekonstruktion von Objekten der chemischen Industrie behindern. Dissertation A, Technische
Hochschule Leipzig.
Ustinovichius, L.; Zavadskas, E. K.; Podvezko, V. 2007. Application of a quantitative multi-criteria decision
making (MCDM-1) approach to the analysis of investment in construction, Control and Cybernetics
36(1): 251–268.
Zagorskas, J.; Turskis, Z. 2006. Multi-attribute model for estimation of retail centres infuence on the
city structure. Kaunas city case study, Technological and Economic Development of Economy 12(4):
347–352.
Zavadskas, E. K.; Peldschus, F.; Kaklauskas, A. 1994. Multiple criteria evaluation of projects in contruction.
Vilnius: Technika. ISBN 9986-05-046-4.
Zavadskas, E. K.; Kaklauskas, A. 2007. Mehrzielselektion für Entscheidungen im Bauwessen [Multi-attribute
decisions in construction]. Fraunhofer IRB Verlag (in German). ISBN 978-3-8167-7203-3.
Zavadskas, E. K.; Peldschus, F.; Ustinovičius, L.; Turskis, Z. 2004. Lošimų teorija statybos technologi-
joje ir vadyboje [Game Teory in Building Technology and Management]. Vilnius: Technika. 2004.
ISBN 9986-05-700-0.
Zavadskas, E. K.; Sivilevičius, H. 2005. Metody vyberu racionalnych variantov stavebnej vyroby, Project-
Stav ba 16(4): 21–24 (in Slovak).
Zavadskas, E. K.; Turskis, Z. 2008. A new logarithmic normalization method in games theory, Informatica
19(2): 303–314.
Zavadskas, E. K.; Turskis, Z.; Vilutienė, T. 2005. Simulation of multi-criteria selection of buildings’ main-
tenance contractor using the game theory, Computer Modelling and New Technologies 9(2): 7–16.
Zavadskas, E. K.; Ustinovičius, L.; Peldschus, F. 2003. Development of sofware for multiple criteria
evaluation, Informatica 14(2): 259–272.
Zavadskas, E. K.; Ustinovičius, L.; Turskis, Z.; Peldschus, F.; Messing, D. 2002. Multiple criteria evaluation
program for construction solution, Journal of Civil Engineering and Management 8(3): 184–191.
545 Technological and Economic Development of Economy, 2008, 14(4): 531–545
LOŠIMŲ TEORIJOS TAIKYMO PATIRTIS STATYBOS VADYBOJE
F. Peldschus
Santrauka
Lošimų teorija teikia matematinių sprendimų konfiktinėse situacijose. Straipsnyje pateikta daug
ekonominių problemų sprendimo pavyzdžių, sukurtų statybos valdymo problemų sprendimo metodikų.
Atlikta šių tyrimų apžvalga, surinktos minėtų inžinerinių problemų sprendimo strategijos. Pateikiami
konkretūs teorijos taikymo pavyzdžiai dabarties sąlygomis ir ateityje. Aptariami „dviejų asmenų nulinės
sumos“ lošimų sprendiniai, taip pat neapibrėžtų aibių teorijos taikymo lošimuose atveju.
Reikšminiai žodžiai: statybos vadyba, variantų atranka, lošimų teorija, „dviejų asmenų nulinės
sumos“ lošimai, lošimai taikant neapibrėžtąsias aibes.
Friedel PELDSCHUS. Professor of building/operating process planning in the area of building industry
of Leipzig University of Applied Sciences; visiting professor to VGTU in Vilnius. Civil engineer, welding
engineer and engineering specialist for data processing. His special feld is the application of the game
theory to compiling optimization solutions in construction. Tis area was also the subject of his Doc-
toral thesis (1972) and his Habilitation thesis (1986). His international popularity resulted in the award
of Doctor honoris causa by Vilnius Gediminas Technical University (1991). He published more than
100 articles in technical periodicals and in 4 books.

532 F. decision criteria considering the totality of all possible results are needed. 1983a. . Zavadskas et al. S2 j for j = 1... Thus.. s2 ) = A s1* . s2 ) = min max A ( s1 . As a solution to this problem. (1) with S1i for i = 1. Extended min–max principle An equilibrium point with mixed strategies is calculated (combination of strategies) as: max min A( s1 . Hollert 2006). Γ = ( S1 . 2. are presented. because for every decision different. a saddle point with pure strategies (one optimal strategy for each player) is obtained as solution – trivial solution. Experience of the game theory application in construction management The investigation of principles is aiming at the elimination of information incompleteness. The application of principles allows to assess decisions with respect to the necessary actions and to give an orientation for selecting them. A ) . that have been solved by this theory. i j i j ( ) (3) . i j j i (2) If α = β = γ . Peldschus. s2* = ν . These games are two-person zero-sum games (Peldschus and Zavadskas 1997.. As it is not known beforehand. the application of the Game Theory is proposed. The Game Theory deals with the representation of conflict situations and their resolution. β = min max aij . and the pay-off function for players I and II: S21 S11 a11 S12 a21   S1m am1 S22 a12 a22  S2n  a1n  a2 n am 2  amn For these games ideally a saddle-point solution (simple min-max principle) or a strategy combination (extended min-max principle) is obtained. n strategies of player II. 2004. which result is actually going to occur. Peldschus 1986) are used.. it allows choosing the most favourable one out of a set of different behaviours.. m strategies of player I. mutually exclusive results can be obtained. Simple min–max principle α = max min aij . Application examples for problems of construction operation. S2 . A simple assessment for all possible results of decisions is not always sufficient. . Theory of the matrix games For solving engineering problems matrix games (Peldschus et al.

M = VAR (Xm).1. 2004. K n } are calculated and added to the decision matrix: VAR. For describing the problem the variants are assigned to the strategies of player I and the criteria to the strategies of player II. this can be written as: VAR. .2 = VAR (X2).. 2001. (2002. a distinction can be made between linear and non-linear transformations (Peldschus 2007a. . 2006. . VAR. Antuchevičienė et al. (4) When generalised. Variants with one variable For the selection of variants.Technological and Economic Development of Economy.M = VAR (Xm) K1 K 2 a11 a12 a21 a22   am1 am 2      Kn a1n a2 n  amn .1 = VAR (X1).2 = VAR (X2)  VAR. Selection of variants The main application of matrix games is the selection of variants.1.. For the pay-off function dimensionless evaluation numbers are used in simple cases.i = VAR (Xi).1. Ginevičius 2008. 2007a. 14(4): 531–545 533 Further specific solution methods for games against nature were discussed in detail in Zavadskas et al. ~0]. Zavadskas and Turskis 2008). 2005. Zavadskas et al. Zagorskas and Turskis 2006. their effectiveness is not comparable. The variants are grouped for solving of the problem.. 1983b. Zavadskas and Kaklauskas 2007. Zavadskas and Kaklauskas 2007. Zavadskas and Turskis 2008). Zavadskas and Sivilevičius (2005). 2007. Labelling the variable with X the following variants are given: VAR. Applications 2. 1994. 2008. VAR. Migilinskas 2003.1 = VAR (X1) VAR. 2007. (5) (6) For these variants the results aij = ai ( X i ) with respect to the criteria K j ∈ { K1 . Generally. Depending on the kind of problem. which is a problem of multi-criteria decisions (Peldschus et al. Ustinovichius et al. Such numbers describe the situation only coarsely. m variants and n criteria can be used. 0] or [1. Brauers and Zavadskas 2007. there are several options for transforming the characteristic values. In order to compare the characteristic values they are mapped on the interval [1. It is therefore sensible to use real characteristic values. 2004). 2. 2003. As such values have different dimensions. Meszek.

The following variants for the 6-lane expansion of a motorway in Thuringia were analysed: Variant 1 – Reconstruction with a concrete surface. modified axis position and gradient.i = VAR (Xi) as strategies of player I. identical axis position and identical gradient.62 59. This includes the refurbishment of damaged motorway segments and the expansion from 4 to 6 lanes. length of the route.02 10. Experience of the game theory application in construction management With the variants VAR. the upgrading of the existing network is of a particular importance. Variant 2 – Reconstruction with an asphalt surface.83 59. The following values were used: Table 1. Example: Multiple-criteria analysis in the construction of motorways The steady growth of road traffic in Germany requires the extension of roads and motorways network.49 12. modified axis position and identical gradient. K n } as strategies of player II and the transformed values aij = ai ( X i ) as pay-off function. modified axis position and gradient. construction period. Peldschus. noise protection originate from operational documents. durability. statistical data were collected in a diploma thesis.90 Durability [years] (max) 30 20 27 18 24 16 Noise level [db(A)] (min) 62.10 10. The values for the criteria costs. In addition to the construction of new routes. For the durability. construction period. Variant 3 – Reconstruction with a concrete surface.61 59. It was aimed at finding the optimal solution that fulfils the mentioned criteria best.37 11. . The selection was made according to the criteria: construction costs. environmental protection and economic efficiency. Variant 6 – Reconstruction with an asphalt surface. Variant 5 – Reconstruction with a concrete surface. the criteria K j ∈ { K1 . identical axis position and identical gradient. Variant 4 – Reconstruction with an asphalt surface. Multiple-criteria analysis in the construction of motorways Costs [mln €] (min) 12.98 11.62 62.61 62. modified axis position and identical gradient.534 F.83 Length of route [m] (min) 1088 1088 992 992 998 998 Construction period [days] (min) 761 746 669 654 700 685 Variant 1 2 3 4 5 6 . a solution as two-person zero-sum game is possible.

m variants and n criteria can be considered. 0.. The values of Table 1 are mapped onto a dimensionless interval using the non-linear transformation (Peldschus):  min aij  bij =  i  aij    aij bij =   max aij i     if min aij is advantageous. M = ∏ Zi . VAR. . Other criteria are of minor importance (Peldschus 2005). 2005) are examples of game theory applying in construction. The variant i is derived from i = ( (Vk − 1) Z k −1 + ν k −1 − 1) Z k − 2 +  + V2 − 1 Z1 + V1 . ... The game-theory equilibrium is formed by the criteria durability and construction period.612 . the selection of wall structures in residential construction (Peldschus et al. 0 ) and S₂ = ( 0 . . X12 . X and the possible values for X l being 1 vl ∈ {1.. 1983b).Technological and Economic Development of Economy.. zl } . 0 . The values of the pay-off function are obtained by transformation to dimensionless values.2 = 1 2 k 1 2 k VAR ( X 2 . this  decision matrix is transformed into a two-person zero-sum game by considering the variants  VAR.254. 14(4): 531–545 535 The numerical solution was performed by the software LEVI which has been developed in a co-operation with HTWK Leipzig and VGTU Vilnius. 0. the following variants are obtained: VAR. 0. With the variants being X . 0 . the selection of buildings’ maintenance contractor (Zavadskas et al. . . Variant 3 gets the highest rating and can therefore be declared as the optimal variant.. confirms this result. X1 ). the estimation of retail centres influence on the city structure (Zagorskas and Turskis 2006). VAR. the following variants are obtained For these variants the results  aij  for the criteria  K j ∈ { K1 . i   2    if max aij is advantageous.746 . . 2008. 0 .1 = VAR ( X1 . X1 .. X Z 2 . Similarly to the formation of variants with one variable.i as strategies of player I and the criteria  K j as strategies of player II. i   3 (7) The solution is made according to the extended min-max principle formula (3). i =1 k ( ) (8) (9) In total.M = ( X Z 1 . . there is the possibility to include several variables in the formation of variants. The formation of variants is performed for all combinations of the 1 k considered variables. K n } are calculated and  added to the decision matrix. 0 . X Zk ). X1k ). treating all criteria as equally important. The calculation of the optimal number of partial assembly lines for work-cycle production (Peldschus 1975).388 ). A comparative calculation with the Laplace criterion. Variants with several variables If the problem cannot be addressed to by a formation of variants with one variable.. 0 . The obtained solution vectors are: S₁ = ( 0.

material consumption planning and structural parameters. A pay-off function has to be determined for both according to the game theory. taking into account the frequency of the different constructions. The uncertainties of the analysis of the building conditions are forming the opposing side. In general. the intended productivity cannot be achieved. With t ∈ {1. 2. As a result. (10) Adding all matrices At′ and dividing by the number of constructions. reserves have been foreseen by construction companies. is obtained as a result.. Calculation of correction values Providing informative numbers for planning is an important problem in construction preparation. then the capacity of the dimensioned system is exceeded. The problem is very obvious in the reconstruction of buildings. however. e} constructions. They can influence the capacity assignment. the matrices At are obtained.. the weighted matrix A* is obtained: A* = t =1 e ∑ At′ ∑ gt e . When the performance is to be increased due to deviations from the planning. problems for which differences between planned and actual efforts are found should be interpreted in this context. Peldschus. In all other cases there is a diminution by a quotient smaller than 1. however. (11) t =1 After transformation to dimensionless values this weighted matrix A* can be used as pay-off function for a two-person zero-sum game.2.. Therefore attempts were made to determine this variable mathematically. the target is reached. When the output is to be decreased. which influence planning and lead to differences between planned and actual efforts.536 F. the matrices for each construction can be multiplied with their frequency. An increase of the planning efforts can only partly compensate these uncertainties. Experience of the game theory application in construction management Consideration of the frequency of the variants If the calculation is to be made for a serial production with different constructions. A production variant. In order to compensate for differences between planned and actual efforts. In this case the value of the pay-off function is set to 1. With the frequency gt this gives: At′ = gt × At . Such analyses are. These values (parameters) are determined by analyses. These reserves are based on estimations. associated with uncertainties. If the amount of the reserve equals the performed construction efforts.. the amount of the reserve is considered as a planning strategy. .1. which can. be not satisfactory. This method has been applied to dimension of the technological line for producing pre-fabricated concrete parts (Altmann and Peldschus 1987). A lack of labour or material can be given as an example.

064 0.186 0.090 0.0 –0.4 0.1. 1+ Z j aij = 1+ Z j 1 + Pi for Pi  Z j .720 0.057 0.3 0.073 0. This.5 0. It is solved according to the Bayes criterion (Peldschus 1986).2 0. Table 2.2 0.1 0.077 0.200 0.073 0.072 0.072 0.073 0.074 0.2 –0.112 0.200 0. As indicated in Table 2.186 0. Pi – planning reserve. in this case the range was set from a shortfall of – 20% to an exceedance of 50% based on statistical investigations..123 0.171 0.080 0. the required range is defined.120 0.063 0.083 0.106 0.080 0.    n  n   S1* =  S1i / S1i ∩ max  ∑ q j aij  ∩ ∑ q j = 1 .185 0.128 0.108 0.114 0.067 0.3.4 0.065 0.104 0..087 0.5 0.058 0.120 0.053 0..064 0. n – number of difference classes (strategies of player II).053 0.068 0. .067 0. 2.169 0. if single evaluations for the defined criteria are available and if the evaluation of the interrelation cannot be performed .067 0.096 0.Technological and Economic Development of Economy.080 0.1 –0.173 0.074 0. 2008.187 0.157 0. n. i j =1 j =1       (13) Example: Plumbing in reconstruction of buildings in Leipzig For the numerical calculation. Complex evaluation The complex assessment is feasible with a game-theory model.059 0.060 0. Plumbing in reconstruction of buildings in Leipzig Zj 0.138 0. m – number of the analysed strategies (strategies of player I).154 0.3 0.080 0.2 0. 14(4): 531–545 537 aij = 1 + Pi for Pi ≤ Z j .120 –0.080 0..064 Pi The calculated planning reserve is 20%. Z j – difference.1 0.1 0.088 0.106 0..064 0. . the two-person zero-sum game is defined.098 0. Such correction values may be also calculated for other crafts (Höher and Peldschus 1981).0 0. (12) and i = 1.073 0. j = 1.068 0.143 0. m.096 0.062 0.

34 68.12 0.16 33. Table 3. the addition would distort the statement.4 161. Thus.09 0. Example: Multi-criteria decision support in facilities management Nine comparable buildings were chosen for this survey.5 155. 2004.00 153.9 133. A weighting of the analysed variants in terms of a global goal is obtained with the gametheory model. These buildings are characterized according to water consumption [m³/m²].06 0. 2002. heat requirement [kWh/m²] and electricity consumption [kWh/m²].0 50. 2002.07 0. Multi-criteria decision support in facilities management Location Air conditioning Yes No No No No No Yes Yes Yes Automatic doors Elevator Consumption of electricity [kWh/m2] 126.18 The numerical processing is done with the help of the computer program LEVI 3. because prices for heat requirement and electricity consumption are not the same.9 Kind of heating District heating District heating Gas Gas Gas Gas Gas Gas Gas Heat consumption [kWh/m2] 39.09 0.31 m3 water/m2 Building1 Building2 Building3 Building4 Building5 Building6 Building7 Building8 Building9 Yes No No Yes No No Yes Yes Yes No No No No No No No No No 0. The game value may be understood as grade of effectiveness with respect to the global goal and may be assigned to a defined classification.40 101.64 119.0 (Peldschus et al.08 0.76 44.0 93.7 125. But their values cannot be summed up. Zavad- . This result may be interpreted as a ranking and it is valid in the sense of shares. As non-monetary criteria.14 70.2 130. Zavadskas et al. Heat requirement and electricity consumption have got the same unit of measurement.13 0. In the simplest case it can be defined by evaluation numbers (point system) or in case of a more detailed input by effective values. The quality of the results depends on the validity of the pay-off function.538 F.08 115. Zavadskas and Sivilevičius 2005.12 0.1 104. The consumption costs also are not relevant because of the partly considerable price difference in various regions. outliers have not been considered. Experience of the game theory application in construction management by summation or by multiplication due to lacking validity. Peldschus. exemplary elements of use were considered which express customer and employee friendliness.

{ ( x. The value can be used for comparison between several groups of buildings (Peldschus and Reichelt 2002). There. ~0]. With regard to criteria. and elevator with 40% respectively. the interval [1. Therefore a transformation is performed to an interval with dimensionless values. Fuzzy games 3. A fuzzy set A in X is a set of ordered pairs. – degree of the engagements of x for a target. Thereby it is feasible to develop numerical solutions to problems which cannot be defined exactly. A= Hereby µ A ( x ) is the degree of association of X with the fuzzy set A. The entirety of the values µ A ( x ) for all potential elements x of X constitutes the association function. 2008. 1].0 allows the following transformations: normalization of vectors. and can be rated as quality level of the group of buildings. (14) . Of these. 14(4): 531–545 539 skas and Turskis 2008) jointly developed by VGTU Vilnius and HTWK Leipzig. The computer program LEVI 3. µ A ( x )) / x ∈ X } . – degree of compatibility of x with a given characteristic. relative variance and non-linear transformation. the assessment of noise control measures (Peldschus 1986) and the sensitivity analysis in land evaluation (Meszek 2004). Because of their functional course. For the task in question the non-linear transformation (formula 7) was chosen. equipment and furnishing are not included in the solution. The association function  µ A ( x ) can be interpreted as: – degree of the contentedness of x in A. Therefore both buildings can be considered as approximately equal.434. The calculated value of the game is υ = 0. the equilibrium point arises from the effect of the criteria water consumption. With values for the range ≤ 2 min the other transformations are only applicable to a certain extent or not at all. 3. – degree of association of x with A. air conditioning/ventilation was taken into account with 50%. Because of the different dimensions or statements. the values in the matrix cannot be compared. 0 ≤ µ A ( x ) ≤ 1 . the difference to building 5 is insignificant. This method has also been applied to assessing the obstructions of construction works in the reconstructed buildings of the chemical industry (Simon 1982). heat consumption and electricity consumption. automatic doors with 10%. Generally it is oriented to the interval [0. building 1 yields the best valuation. – degree of the possibility of incidence of x.Technological and Economic Development of Economy. This association function µ A ( x ) can reach all values of non-negative real numbers. This decision is justified because of the values range of the criteria examined. However. Theory of fuzzy games New approaches can be developed by applying the theory of fuzzy sets to the game theory. The game-theoretical equilibrium point (formula 3) of the game is formed from the evaluation of the buildings 1.1. linear transformation. 5 and 9.

5). if xD is near a0 or xM . . . . if x has the maximum degree of association to A. 1) and the boundary conditions µ′A ( x0 ) = µ′A ( xM ) = 0 and µ′′ ( xD ) = 0 two third-grade polynomial are obtained. which is usually the case: x − xD (17) −1 + 2 ≤ M ≤ 1+ 2 . 3 2 ( G1 ) A x0 + B x0 + Cx0 + D = 0 .5.5 . 0). xD − x0 Complications only arise. For a fuzzy decision association function µ A ( x ) indicates to which grade each element x complies with the given requirements. and the selection of the nodes. which is defined by the requirements to µ A ( x ) of continuity. 1] and it is also monotone in xi . ( xD.540 F. Experience of the game theory application in construction management – degree of comprehension of x in an issue.. (16) The equation system (G1 ). x0 ≤ x ≤ xD . 2 2 ( G7 ) 3A xD + 2B xD + C − 3E xD − 2F xD − G = 0 . xM ≤ x. Thus for xm > x0 the following approach is obtained: 0   3 2  Ax + Bx + Cx + D µA ( x)  3 2  Ex + Fx + Gx + H  1  for for for for x ≤ x0 . With the nodes ( x0. These are attached continuously A in xD in a way that they can be differentiated twice. 0.. (xM . 3 2 ( G4 ) E xM + F xM + GxM + H = 1. existence of the first and second derivatives.. Functions of an S-shape are more suitable for solving engineering problems. H) are calculated from an equation system (14). Hence it has to be re-defined for every problem. The interpolating cubic spline function has proven to be suitable for practical problems of multi-criteria decisions (Omran 1988). – degree of concordance of x with a given statement. In order that µ A ( x ) ⊂ [0. (G8 ) is not singular and possesses a unique solution. An element x ∈ X represents an optimal fuzzy decision.. 2 ( G5 ) 3A x0 + 2B x0 + C = 0 . Piecewise linear functions are often used. Peldschus. ( G8 ) 6A xD + 2B − 6E xD − 2F = 0 .. xD ≤ x ≤ xM . Different concepts for determining the association function are known.. the following condition must be fulfilled. 3 2 ( G2 ) A xD + B xD + CxD + D = 0. 3 2 ( G3 ) E xD + F xD + GxD + H = 0. (15) The coefficients (A. 2 ( G6 ) 3E xM + 2F xM + G = 0 . The interpretation of the association function is task dependent.

As compensation between the criteria is admitted.. strategies of player II... association function for the pay-off function j = 1. are known. For each criterion the association function (15) is calculated. This first gives a basic matrix for the game Γµ to be solved. pay-off function j = 1... The set of the strategies and the set of the criteria.. .( S2 j . association function for the strategies of player II with respect to strategies of player I .. a summation is done by calculating the arithmetic mean (Laplace-criterion).. For the solution. can be described as follows:  Γµ = ( S1i ..Technological and Economic Development of Economy. A widely accepted approach is based on Bellmann and Zadeh (1970). the norm values are transferred to association values. { } (18) with S1i for i = 1. 14(4): 531–545 541 For the combination of the theory of matrix games and the theory of fuzzy sets a threestep model was proposed (Omran 1988). . . adopted from the classic game theory... m strategies of player I .. For the average of the fuzzy sets Z and R the association function is defined point wise by the operator ' µij ( x) = min [µ Z ( x). . (20) As a result. n aij for i = 1.... In contrast to the first step. .e.. .... . Also. The mapping of sets is done in matrix format. n S2 j for j = 1. n... for each strategy of player I an association value for several criteria is obtained.(aij . it is an intersection of the strategies  of players I and II. The transition from game Γ to game Γµ is done in 3 steps: 1st step On the set of strategies of player I a fuzzy set is defined. µ R ( x) ] . which contains partly or wholly fuzzy information. µij ) . the actual fuzzy game matrix is obtained. association function for the strategies of player I. the min-maxprinciple. n µij for j = 1. i. L i =1 nd 2 step In the second step the strategies of player II are considered. 1 L (19) µ1i = ∑ µil . is used. . The fuzzy game Γµ. on the set of strategies of player II fuzzy sets are defined and association values are calculated according to (15). Thus.. n. m µ1i for i = 1. 3rd step The third step is combination of step 1 and 2. Thus. µij ). m  µij for i = 1. in which the fuzzy decision is defined as the intersection of the fuzzy sets for the fuzzy goals Z and the fuzzy set of the restrictions R.. 2008.. this matrix is interpreted with respect to the game theory. . µ1i ). which quantitatively describe the set of the strategies.

the required capacity etc. . which could be discussed for the problem to be solved. is obtained as a result. so that the application of a compensatory parameter (17) is appropriate. are shown. lack of information and fuzziness. The parameters represent the internal influences and relating to the construction phase. A matrix. They have the character of prognosis and represent the utilisation phase. are considered. The solution of fuzzy games forms a new quality of decisions.2. 4. which describes the fuzzy game. The interrelation of the strategies of player II with the strategies of player I are described with the minimum operator (18). etc. arise. which represent a high degree of complexity. Experience of the game theory application in construction management 3. These methods differ substantially from those applied so far. This concept was developed for considering internal and external influences. economical fields. The strategies of player II contain the influences from the use of the building. The result of these investigations is the selection of the optimal variant in consideration of several criteria and the compliance with practical conditions associated uncertainty. which are constituted by the consideration of several. All real construction variants. the operational and energy costs. Results The game theory applications show new possibilities of the solution of conflict situations in construction. which are achieved by the application of mathematical methods. the construction costs. Applications The developed algorithm for fuzzy matrix games is a fuzzy concept for multi-criteria decisions. objectives. association values are calculated according to (13).542 F. They represent the external influences. With fuzzy games a lot of new solution possibilities for problems from technical fields. Thus an algorithm is available to aggregate also quality characteristic with a hierarchical structure. i.e. the construction time. These serve the attempt for gaining maximum benefit in conducting construction works. partly contradictory. Internal influences have the character of experiences and they apply until the system comes in use. even in case of a number of comparable situations. the following examples can be given: the required space. As specific parameters. External influences describe a new quality. Therefore approaches. The solution is found using the min-max principle (3). For practical applications the construction variants become the strategies of player I. Specific investigations have been performed for variant evaluation of cable routes in the chemical industry and for water supply systems (Peldschus and Zavadskas 2005) as well as for multiple-criteria decisions in risk management (Reichelt and Peldschus 2005). These construction variants are evaluated with respect to several parameters. It is known which difficulties are associated with such a task. Compensation between the parameters is admitted. the amortisation. For these criteria values. their effective share is expressed. in the construction or production phase. It has a strategical character. With the calculation of the association values for the strategies of player I. hereby admitting different phases. These include: the resistance against failure of the construction. Peldschus. investment planning.

Migilinskas. E.. J. K. Peldschus. 1988. which need to be considered in the solution as well as in the formalisation of the pay-off function. Ginevičius. 2007. 231–233. Zur Anwendung der Theorie der Spiele für Aufgaben der Bautechnologie. J. E. Multiple-Criteria Analysis in the Construction of Motorways. A. 119 p. 2006. Berichte Heft 1. Rechnergestützte Optimierung bei der technologischen Bemessung von Fertigteillinien für die Betonelementeherstellung.. F. P. 2005. Meszek. 429 S. 1975. 2001. and Zadeh. 1970. W. Meszek. Entscheidungsunterstützende Methoden in der bautechnologischen Vorbereitung unter Berücksichtigung von unscharfen Spielen – Fuzzy Games. F. Zavadskas. Here a clear and unambiguous definition of the problem is necessary. Every formal description of a matrix game exhibits particularities. 46–47. W. Einführung in die Spieltheorie. Dissertation A. proven. D. Antuchevičienė. Hollert. It was. Technological and Economic Development of Economy 12(4): 263–268. K. . M. The purpose of the results is always the decisive element. Bellman. Dissertation B. Reihe B Heft 4. Weimar. International Journal of Management and Decision Making 8(5–6): 575–585. 2008. Technische Hochschule Leipzig. K. Brauers.. 14(4): 531–545 543 The application of the theory of matrix games may seem simple. Technological and Economic Development of Economy 7(2): 62–68. This issue becomes more and more important in the light of increasing complexity of construction preparation. 1987. also complex. Internationaler Kongress über die Anwendung der Mathematik in den Ingenieuwissenschaften. Omran. W. Peldschus.Technological and Economic Development of Economy. F. in 11.. Peldschus. Wissenschaftliche Zeitschrift der Hochschule für Bauwesen Leipzig 1/75: 53–55. The influence of normalization methods selection in construction including game theory adaptation. Normalization of quantities of various dimensions. G. 2006. that every real conflict situation is difficult to be analysed due to many. R. and the need for objectiveness gains new dimensions. 2007. Höher. Meszek. Estimation of land plots using game theory. J. 2004. Z. E. Normalisation in decision-making methods. L. in 9. Weimar. Die optimale Anzahl der Teiltaktstraßen für die Fließfertigung. Economics and Management 9(1): 79–86. 1986. Springer Verlag Berlin. Modelling renewal of construction objects applying methods of the game theory. R. Decision-making in a fuzzy environment.. Technological and Economic Development of Economy 10(1): 40–46. influences. Peldschus. 2008. Wiss Zeitschrift der Hochschule für Architektur und Bauwesen Weimar. Journal of Business. Technological and Economic Development of Economy 11(1): 32–35. Editorial. Peldschus. however. Eine Methode zur Verbesserung von Rekonstruktionsvorhaben im Wohnungsbau. Internationaler Kongress über die Anwendung der Mathematik in den Ingenieuwissenschaften. Uncertainty phenomenon on property valuation. Zavadskas. The application of the theory of matrix games will contribute to the improvement of the objectiveness in construction preparation. Turskis. Technological and Economic Development of Economy 9(2): 73–79. W. 137 p. References Altmann. F. Management Science 17(4):141–164. Technische Hochschule Leipzig. International Journal of Management and Decision Making 8(5-6): 441–444. F. 1981. Usage theory of the plays for the plot value inquiry. 2003.

Games-theory solutions in construction operation.... K. Metody vyberu racionalnych variantov stavebnej vyroby. E. 2006. L. in Proceedings Volume of the Meeting of the European Working Group of Multiple Criteria Decision Aiding. Zavadskas. Bauplanung–Bautechnik 37(4): 173–175. Peldschus. 134 p. F. 2004.. E. Peldschus. Mehrzielselektion für Entscheidungen im Bauwessen [Multi-attribute decisions in construction].. K. Eine Methode zur Erfassung. 1982. K. Peldschus. 2003. Development of software for multiple criteria evaluation.. Turskis. Turskis.). Application of a quantitative multi-criteria decision making (MCDM-1) approach to the analysis of investment in construction. Turskis.0 – Multiple Criteria Evaluation Program under uncertainty. K. LEVI 3. May 16–18. 6: 159–173. Zavadskas. Technological and Economic Development of Economy 8(1): 3–12. Control and Cybernetics 36(1): 251–268. Reichelt. D.. 2004... E. Technologische Entscheidungen bei der Berücksichtigung mehrerer Ziehle. L. Zavadskas.. K. Zavadskas. K. Skibniewski. Kaklauskas. Informatica 14(2): 259–272. Kaunas city case study. Reichelt. F. H. Ustinovičius. Ustinovičius. F. 2005. L. E. E. L. Multiple criteria evaluation program for construction solution. F.. 2007. Zavadskas (eds. K. Peldschus. K. 2002. Multicriteria Decision Support in Facility Management. ISBN 9986-05-046-4. ISBN 978-3-8167-7203-3. Zagorskas.. Informatica 16(1): 107–120. Z. Experience of the game theory application in construction management Peldschus. F.. Matriciniai lošimai statybos technologijoje ir vadyboje [Matrix Games in Building Technology and Management].. In Modern Building Materials.. E. Vilnius. Lithuania. D. Turskis. . K. E. 2002. E. editor European Commission Joint Research Centre. Ustinovičius. E. Peldschus. 1983. F. Foundations of Civil and Environmental Engineering. Peldschus. Informatica 19(2): 303–314. Zavadskas... Vilnius: Technika. K. Zavadskas. Vilnius: Technika. Journal of Civil Engineering and Management 8(3): 184–191. Z. die den Bauablauf der Rekonstruktion von Objekten der chemischen Industrie behindern. Selected Papers. Z. K. Z. Fuzzy Matrix Games Multi-Criteria Model for Decision-Making in Engineering. ISBN 9986-05-700-0. Ustinovičius. Peldschus. The Application of Multi-Criteria Decision Analysis (MCDA) in Risk Management of Civil and Environmental Engineering Projects. Peldschus. Bewertung und Klassifizierung von Einflüssen. Vainiunas. Technische Hochschule Leipzig. Zavadskas. 2007. Z. Zavadskas. Computer Modelling and New Technologies 9(2): 7–16. V. 2007a. F. K. E. Zavadskas. Peldschus.. Simon. E. The 9th International Conference. F. The effectiveness of assessments in multiple criteria decisions. Zavadskas. A. E. Multi-attribute model for estimation of retail centres influence on the city structure. A. A new logarithmic normalization method in games theory.. 2005. Vilnius: Technika. Zavadskas. ISBN 9986-05-332-3. Z. Podvezko. Vilutienė.. F. B. F. Ustinovichius. Peldschus.. E. Vaigauskas. 2008. Turskis. Poznan University of Technology. Peldschus. Lošimų teorija statybos technologijoje ir vadyboje [Game Theory in Building Technology and Management]. E. Peldschus. Zavadskas. Kaklauskas. 348–353. ProjectStav ba 16(4): 21–24 (in Slovak). 2007b. Simulation of multi-criteria selection of buildings’ maintenance contractor using the game theory. in M.. L. Messing. K. 2005. L. International Journal of Management and Decision Making 8(5–6): 519–526.544 F. F. 1994. 1997. Structures and Techniques.. 2002. J. Coimbra PT.. Zavadskas. Fraunhofer IRB Verlag (in German).. Dissertation A.. B. J. P.. E. Multiple criteria evaluation of projects in contruction. Turskis. Technological and Economic Development of Economy 12(4): 347–352. Messing.. 2007. 2005. K. F. Sivilevičius.. T.

Aptariami „dviejų asmenų nulinės sumos“ lošimų sprendiniai. Atlikta šių tyrimų apžvalga. Civil engineer. Professor of building/operating process planning in the area of building industry of Leipzig University of Applied Sciences. lošimų teorija. welding engineer and engineering specialist for data processing. Reikšminiai žodžiai: statybos vadyba. Peldschus Santrauka 545 Lošimų teorija teikia matematinių sprendimų konfliktinėse situacijose.Technological and Economic Development of Economy. 14(4): 531–545 LOŠIMŲ TEORIJOS TAIKYMO PATIRTIS STATYBOS VADYBOJE F. „dviejų asmenų nulinės sumos“ lošimai. He published more than 100 articles in technical periodicals and in 4 books. His special field is the application of the game theory to compiling optimization solutions in construction. This area was also the subject of his Doctoral thesis (1972) and his Habilitation thesis (1986). Pateikiami konkretūs teorijos taikymo pavyzdžiai dabarties sąlygomis ir ateityje. Friedel PELDSCHUS. visiting professor to VGTU in Vilnius. . Straipsnyje pateikta daug ekonominių problemų sprendimo pavyzdžių. lošimai taikant neapibrėžtąsias aibes. surinktos minėtų inžinerinių problemų sprendimo strategijos. His international popularity resulted in the award of Doctor honoris causa by Vilnius Gediminas Technical University (1991). sukurtų statybos valdymo problemų sprendimo metodikų. 2008. variantų atranka. taip pat neapibrėžtų aibių teorijos taikymo lošimuose atveju.

Sign up to vote on this title
UsefulNot useful