Computers and Operations Research 82 (2017) 180–188

Contents lists available at ScienceDirect

Computers and Operations Research

journal homepage: www.elsevier.com/locate/cor

Improving polynomial estimation of the Shapley value by stratified

random sampling with optimum allocation
Javier Castro a,∗, Daniel Gómez a, Elisenda Molina b, Juan Tejada c
a
Departamento de Estadística e I.O. III. Facultad de Estudios Estadísticos, Universidad Complutense de Madrid, Av. Puerta de Hierro s/n. 28040, Madrid,
Spain
b
Departamento de Estadística, Universidad Carlos III de Madrid, Calle Madrid, 126. 28903, Getafe, Madrid, Spain
c
Instituto de Matemática Interdisciplinar (IMI), Departamento de Estadística e Investigación Operativa, Facultad de Ciencias Matemáticas, Universidad
Complutense de Madrid. Plaza de Ciencias 3. 28040, Madrid Spain

a r t i c l e i n f o a b s t r a c t

Article history: In this paper, we propose a refinement of the polynomial method based on sampling theory proposed
Received 5 April 2016 by Castro et al. (2009) to estimate the Shapley value for cooperative games. In addition to analyzing
Revised 26 January 2017
the variance of the previously proposed estimation method, we employ stratified random sampling with
Accepted 29 January 2017
optimum allocation in order to reduce the variance. We examine some desirable statistical features of the
Available online 31 January 2017
stratified approach and provide some computational results by analyzing the gains due to stratification,
Keywords: which are around 30% on average and more than 80% in the best case.
Computer science
© 2017 Elsevier Ltd. All rights reserved.
Game theory
Operations research
Shapley value
Stratified sampling algorithm

1. Introduction
One of the most important one-point solution concepts in the
framework of coalitional games with side payments is the Shapley
value (Shapley, 1953), which proposes the allocation of resources
in multiperson interactions by considering the power of the play-
ers during their various opportunities for cooperation. Since the
pioneering work by Shapley, the Shapley value has been studied
widely from a theoretical viewpoint. Many studies have also fo-
cused on the potential applications of the Shapley value to specific
problems (e.g., see Lucchetti et al., 2010; Moretti, 2010). However,
since computing the Shapley value is an NP-complete problem (see
Deng and Papadimitriou, 1994; Fernández et al., 2002, or Faigle
and Kern, 1992, for more details), the problem of its calculation
must be addressed before it can employed as a useful tool in real
situations.
Two main approaches have been employed to address the cal-
culation of the Shapley value: developing efficient strategies to
compute the Shapley value exactly.
Different approaches have been considered for the exact calcu-
lation of the Shapley value. In order to compute the Shapley value
exactly, some specific features of the problem and/or the game
are assumed. For instance, the Shapley value was obtained in a
polynomial manner for some voting games by Bilbao et al. (20 0 0).
(Granot et al., 2002) developed a polynomial algorithm to obtain
the Shapley value of an extended tree game, where a tree network
serving heterogeneous customers was constructed. (Castro et al.,
2008) computed the Shapley value for an airport game in polyno-
mial time by considering that this value was obtained using the se-
rial cost sharing rule. In addition, Deng and Papadimitriou (1994),
Bolus (2011), Chalkiadakis et al. (2012), and Conitzer and Sand-
holm (2006) developed efficient strategies for computing the Shap-
ley value by finding alternative efficient representations for transfer-
able utility (TU) games. However, although these methods aim to
serve as general tools, their applications are constrained to some
specific classes of games, particularly weighted majority games and
combinatorial optimization games. Moreover, all of these methods
require a specific representation of the game.
By contrast, few studies have focused on the approximation of
the Shapley value. Considering the widespread application of game
theory to real-world problems where exact solutions are often not
possible, then it is necessary to develop algorithms that may facil-
itate this approximation.

∗
Corresponding author. (i) To the best of our knowledge (also see Maleki et al.,
E-mail addresses: jcastroc@estad.ucm.es (J. Castro), dagomez@estad.ucm.es (D. 2014), the first attempt to estimate the Shapley value for
Gómez), elisenda.molina@uc3m.es (E. Molina), jtejada@mat.ucm.es (J. Tejada). a large class of games was the algorithm proposed by

http://dx.doi.org/10.1016/j.cor.2017.01.019
0305-0548/© 2017 Elsevier Ltd. All rights reserved.
 J. Castro et al. / Computers and Operations Research 82 (2017) 180–188
Castro et al. (2009), who suggested the estimation of the
Shapley value (or any semivalue) for cooperative games
based on sampling. The only constraint on this algorithm be-
ing polynomial is the fact that the worth of any coalition
must be computed polynomially. This approach generalizes
the earliest approximation method proposed by Mann and
Shapley (1960) for estimating the power index of a weighted
voting game. In fact, Mann and Shapley (1960) stated that:
“These methods have potential applications beyond the par-
ticular problem solved here.” Moreover, as proposed in the
present paper, Mann and Shapley also considered alterna-
tive sampling methods for reducing the variance of the basic
Monte Carlo sampling approach.
(ii) If an efficient general purpose algorithm is available, then it
will be possible to estimate the Shapley value by refining
the original procedure by Castro et al. in order to exploit the
specific properties of the problem considered. Thus, Maleki
et al. (2014) proposed the use of stratified random sam-
pling to reduce the estimated error in the original method
described by Castro et al. They also considered the same
strata treated in the present paper. However, their allocation
method ignores the specific features of the game, thereby
resulting in an estimation method that improves the origi-
nal method based on simple random sampling, but it gen-
eral performs poorly compared with the refinement that
we propose in the present study. The reader may refer to
Section 6 for a comparative analysis using the cooperative
games analyzed by Castro et al. (2009).
Few studies have developed approximations to estimate the
Shapley value and most of them are problem specific. Fatima et al.
(2006) presented a randomized polynomial method for determin-
ing the approximate Shapley value for weighted voting games and
k-majority voting games. Crama and Leruth (2007) also proposed
the estimation of the Shapley value for a complex voting game
that reflects the structure of shareholdings in complex interlocked
shareholding structures based on a Monte Carlo simulation.
In this paper, we propose a refinement based on stratified ran-
dom sampling, which we use instead of simple random sampling
to estimate the Shapley value. As found with the original algo-
rithm, these refined estimations are polynomial if each marginal
contribution can be calculated in polynomial time. The remainder
of this paper is organized as follows. In Section 2, we provide some
basic concepts and notation. We also describe the ApproShapley al-
gorithm proposed by Castro et al. (2009), which is the basis of our
propose method. In Section 3, we analyze the sources of the vari-
ance in the estimations derived from ApproShapley and we propose
three methods for reducing this variance. First, we propose a gen-
eral method based on stratified random sampling St-ApproShapley,
Shst , and we then determine the optimum allocation of the sam-
ple sizes in the corresponding strata St-ApproShapley-opt, Shst, opt .
We also propose a two-stage estimation method that preserves the
efficiency of the Shapley value, St-ApproShapley-eff, Shst, eff , in the
case where efficiency is a demanding property. In Section 4, we
analyze the estimated error for stratified random sampling with
the optimal allocation approach, and we compare it with the es-
timation method proposed by Maleki et al., which is described in
Section 3. In Section 5, we present the proposed estimation algo-
rithm. Finally, in Section 6, we present some computational results
to show the improvements obtained using the proposed method
for the cooperative games analyzed by Castro et al. (2009). We give
our conclusions in Section 7.
2. Preliminaries
In this section, we introduce some sampling notation and theo-
retical games in order to understand the rest of this paper. We also
181
recall the original sampling algorithm proposed by Castro et al.
(2009) upon which our refinement is based.
A cooperative game in coalitional form with side payments, or
with TU, is an ordered pair (N, v), where N is a finite set of players
and v : 2N → R, with 2N = {S | S ⊆ N}, is a characteristic function on
N that satisfies v(∅ ) = 0. For any coalition S ⊆ N, v(S ) ∈ R is the
worth of coalition S and it represents the reward that coalition S
can obtain by itself if all its members act together. We restrict to
the case of TU games in the sequel, so we refer to them simply as
games. For brevity, throughout this paper, the cardinalities of the
sets (coalitions) N and S are denoted by appropriate small letters n
and s, respectively. In addition, for convenience, we write the sin-
gleton {i} as i if there is no ambiguity. When the worth v(S) repre-
sents the cost that players must be charged, the game is known as
a cost-game and we denote these types of games as (N, c).
Let GN be the vector space of all the games with a fixed player
set N, and identify (N, v ) ∈ GN by its characteristic function v if
there is no ambiguity. A value ϕ is an allocation that associates
with each game (N, v ) ∈ GN a vector ϕ (N, v ) ∈ RN , where ϕi (N, v ) ∈
R represents the value of player i, i ∈ N. Shapley (1953) defines his
value as follows:
 ( n − s − 1 )! s !
Shi (v ) = (v(S ∪ {i} ) − v(S )), i ∈ N. (1)
n!
S⊆N
i∈S
The value Shi (v) of each player, which is a weighted average of his
marginal contributions, allows different interpretations, such as the
payoff that player i receives when the Shapley value is used to pre-
dict the allocation of resources in multiperson interactions, or his
power when averages are used to aggregate the power of players
in their various opportunities for cooperation.
Weber (1988) gave an alternative characterization of the Shap-
ley value in terms of all the possible orders of the players, which
was employed for estimation by Castro et al. (2009) . Let O: N
→ N be a permutation that assigns the player O(k) to each po-
sition k and let π (N) denote the set of all possible permutations
of the player set N. Given a permutation O ∈ π (N), let us de-
note Prei (O) as the set of predecessors of the player i in the or-
der O (i.e., P rei (O ) = {O(1 ), . . . , O(k − 1 )}, if i = O(k )). In this set-
ting, the vector of marginal contributions for a given order O ∈ π (N),
x(O ) = (x(O )i )i∈N is defined as:
x(O )i = v(P rei (O ) ∪ i ) − v(P rei (O )), i ∈ N.
Weber (1988) showed that the Shapley value can be expressed
as the following expectation, where it is assumed that all different
orders have equal probability:
 1
Shi (v ) = x(O )i , i ∈ N. (2)
n!
O∈π ( N )
Thus, since the Shapley value is an expectation, Castro et al.
(2009) proposed its statistical estimation. Next, we describe the al-
gorithm that they proposed (emphApproShapley), which is based
on a unique simple random sampling process for estimating the
Shapley value Shi of all players i ∈ N.
1. The population of the sampling process P is the set of all possi-
ble orders of players, i.e., P = π (N ). Each sampling unit repre-
sents an order O ∈ π (N).
2. The vector parameter under paper is Sh = (Shi )i∈N .
3. The characteristic observed in each sampling unit, O ∈ π (N), is
the vector of the marginal contributions for that order O, i.e.,
x ( O ) = ( x ( O )i )i∈N .
4. The estimate of the parameter Sh, Sh ˆ , is the mean of the
marginal contributions over the sample M, i.e., Sh ˆ = (Sh
ˆ i )i∈N ,
where

ˆi= 1
Sh x(O )i , i ∈ N.
m
O∈M
182 J. Castro et al. / Computers and Operations Research 82 (2017) 180–188

5. Finally, the selection process used to determine the sample M St-ApproShapley-eff, Shst, eff , for the case where efficiency is a de-
will take any order O ∈ π (N) with a probability of n1! . They con- manding property.
sider a sample M as a subset of P × P . . . × P, i.e., a sample that The first stratified random sampling process, which we refer to
  
m times as St-ApproShapley, is defined as follows.
has been obtained with replacement.
1. The population in the sampling process is P and each sampling
unit represents a marginal contribution X(O)i ∈ P.
ˆ can be calculated in polyno-
Note that in order to assure that Sh
2. The parameter under paper is Sh.
mial time, it is assumed that the worth of any coalition S, v(S), can
3. The characteristic observed in each sampling unit is x(O)i .
be calculated in polynomial time. In fact, it is sufficient to assume
4. The strata considered are Pi for all i = 1, . . . , n and  = 1, . . . , n,
that the vector of the marginal contributions of all orders can be
where Pi is the set of all possible marginal contributions in
obtained in polynomial time.
which player i is in position .
We select a random sample Mi of size mi , which is obtained
with replacement from each stratum Pi . Thus, the selection
3. Improving the estimation of the Shapley value: Stratified
process used to determine the sample Mi will take any ele-
random sampling approach
ment of Pi with a probability of (n−11
)! .

In some situations, the ApproShapley algorithm proposed by 5. Let Shi be the mean of the marginal contributions over the
Castro et al. (2009) might not be sufficiently precise for our pur- sample Mi , i.e.
poses. In these cases, a method must be employed to reduce the  1 
variance of the estimations obtained by simple random sampling. A Shi = x ( O )i , ∀i = 1, . . . , n ∀ = 1, . . . , n.
mi
common technique for variance reduction when using Monte Carlo X (O )i ∈Mi

methods is to estimate the expectations from a known population Then, the estimate of the parameter Shi obtained by stratified
by stratification (e.g., see Cochran, 1977 or Lohr, 1999), so we pro- sampling is Shst
i
, where
pose the use of stratified random sampling instead of simple ran-

n
1 
dom sampling to improve the performance of the original method. Shst
i = Sh , ∀i = 1, . . . , n.
Stratification may enhance the estimated precision for the expecta- n i
=1
tions in the whole population if it is possible to divide a heteroge- Note that a procedure for determining the sample size mi
neous population into subpopulations, each of which is internally taken from each stratum Pi must be defined in order to apply
homogeneous. In many cases, one of the sources of the variance the St-ApproShapley algorithm described above. We propose the
in each marginal contribution is the player and the order in which use of stratification with the optimum allocation of mi (Neyman,
this player arrives, so we consider these two sources of variation 1934), which means that the sampling sizes mi in the correspond-
when building our subpopulations. ing strata are selected to minimize the estimation error. The next
For example, let us consider the airport game described in proposition shows that Shst is unbiased, and thus we select the
i
Section 6, where the marginal contributions of larger planes vary sample sizes in order to minimize the variance of the estimator,
substantially depending on whether the plane arrives at the begin- V [Shst ].
i
ning or the end of the corresponding order, where they are inter-
nally homogeneous in each of these positions. This is generally the Proposition 3.1. The estimator Shst
i
is unbiased, i.e., E[Shst
i
] = Shi ,
case for games determined by a graph structure where hub players and its variance is given by:

exist that connect large connected components. In addition, in the
1  σi2 (x(O )i − Shi )2
X (O )i ∈Pi
case of voting games, the marginal contributions can vary from 0 V [Shst
i ] = 2 , where σ =
2
, (3)
to 1 depending on the position in which the voter arrives, and in
n ∈L mi i
( n − 1 )!
i

the case of extreme weights, they tend to be stable for each given and Li ⊆ {2, . . . , n − 1} is the set of the indices of the strata for player
position. i with strictly positive variability and Shi is the expected marginal
Therefore, criteria for determining the strata may comprise the contribution of player i when he occupies the th position:
players and the order in which each player arrives. In this case, 
X (O )i ∈Pi x ( O )i
we define n × n strata, Pi , i ∈ N,  = 1, . . . , n, where the stratum Shi = .
Pi is the set of all the possible marginal contributions in which ( n − 1 )!
player i occupies position . Let us note that the population (from Proof 3.1. The proof is straightforward given the properties of the
which we take random samples) must include all possible marginal sample mean as an estimator of the population mean. Note that
contributions, so we have n · n! possible units. This number (n · in St-ApproShapley, the samples are drawn independently from dif-
n!) is obtained because for each n marginal contributions, we have ferent strata. Moreover, since all the values x(O)i in strata Pi1 are
n! possible orders. The population with size n · n! must be di- v(i) and all the values x(O)i in strata Pin are v(N ) − v(N \ i ), then
vided into non-overlapping subpopulations, Pi for all i = 1, . . . , n σi21 = σin2 = 0 for every player i ∈ N. 
and  = 1, . . . , n, which comprise the whole of the population. Af-
ter the strata have been determined, a random sample is taken According to expression (3), the sum of variance of all play-
n st
from each stratum. ers, i=1 V [Shi ], is minimized when mi is proportional to the
In this paper, we propose three methods for reducing the vari- true variance σi2 in each stratum i = 1, . . . , n and  = 1, . . . , n. Let
ance in the original estimation method (Castro et al., 2009). First, Shst, opt be the estimator of the Shapley value obtained by stratified
we propose a general method based on stratified random sampling random sampling with the optimum allocation, i.e., with mi ∝ σi2 .
(which we refer to as St-ApproShapley and its solution is Shst ). Next, In this case, m → ∞ implies that mi → ∞ for every stratum Pi
we determine the optimum allocation of the sample sizes in the with σi2 > 0. Therefore, the consistency of Shst, opt can be estab-
corresponding strata (we refer to this method as St-ApproShapley- lished as a corollary of Proposition 3.1.
opt and we denote its solution as Shst, opt ). We also describe the
estimation method proposed by Maleki et al. (2014) as a particu- Corollary 3.1. The estimator Shst, opt (m) is consistent in probability,
i.e.,
lar case of St-ApproShapley. Finally, we present a two-stage estima- 
tion method, which preserves the efficiency of the Shapley value, lim P |Shst,opt
i
(m ) − Shi | >  = 0, ∀  > 0 ∀ i = 1, . . . n, (4)
m→∞
 J. Castro et al. / Computers and Operations Research 82 (2017) 180–188 183

where Shst, opt (m) denotes the estimator of Sh obtained by stratified n strata Si0 , Si1 , . . . , Sin−1 defined as Sik = {C ⊆ N \ {i}, |C | = k}. In par-
random sampling with the optimum allocation based on a total sam- ticular, they proposed estimating the average marginal contribu-
 
ple M = ∪ni=1 ∪n=1 Mi of size m = ni=1 n=1 mi . tion of player i in each stratum:
1 
In the case where efficiency is a desirable property, we propose φik = (v(C ∪ {i} ) − v(C )),
an alternative estimation method referred to as St-ApproShapley-eff, |Sik | C∈Sk
i
which preserves the efficiency, reduces the variance in the original
simple random sample estimation approach ApproShapley, and re- k = 0, 1, . . . , n − 1, i = 1, . . . , n,
mains an unbiased estimator of Sh. We propose a two-stage esti- by simple random sampling, and then obtained the final estimation
mation approach by combining the original unbiased and efficient of Shi as the average
ApproShapley approach with the St-ApproShapley with optimum al-
n−1
location approach. 1  ˆk
Let m be the total sample size, and let msim and mst be such
φˆ i = φi
n
k=0
that m = msim + mst . Then, the propose two-stage efficient strati-
fied method is defined as follows. They proposed the calculation of the sample size in each stratum,
as follows. Let
• Stage 1. Efficient estimation based on a simple random sam-
mi ( k + 1 ) 3
2

ple Msim obtained with replacement from the set P = π (N ) of m∗k

i = n−1 , k = 0, . . . , n − 1, i = 1, . . . , n. (6)
j=0 ( j + 1 )
2
sim
all possible orders of N players of size mn , where we use the 3

ˆ.
original ApproShapley method to obtain the estimator Sh Then, allocate mki = m∗k i
samples to stratum Sik , k = 0, . . . , n − 1,
• Stage 2. Perform a St-ApproShapley with optimum allocation us-
st,opt and sequentially increase the value of mki by 1 from k = 0 to (n −
ing a total sample size of mst . Let Shst,opt = (Shi )i∈N be the
1 ) until it exceeds mi , where mi is the total sample size allocated
estimator obtained.
to the estimation of Shi .1 . The reader is referred to Remark 4.2 for
• Estimator. The final estimate of the parameter Sh, Shst, eff , that
the motivation of m∗ki
.
we propose is given by:
Now, each stratum Sik is equivalent to Pik+1 , so the method pro-


n  Sˆh posed by Maleki et al. is a particular case of St-ApproShapley with
( ) , ∀ i ∈ N.
i
Shst,e ff
= Shst,opt + v N − Sh st,opt
(5) allocation
i i
i=1
i v ( N)
 
   m · 3
2

efficiency gap m∗i = n 2

,  = 1, . . . , n, i = 1, . . . , n. (7)
n =1  3
Proposition 3.2. The estimator Shst, eff is an efficient allocation, i.e.,
 4. Analysis of the gain due to stratification
Shst,e
i
ff
= v ( N ).
i∈N In this section, we analyze the precision of the proposed esti-
mation method. In particular, we analyze the gain due to strati-
Proof 3.2. The proof is straightforward given the efficiency of the
fication when we use optimal allocation, i.e., we focus on the St-
estimation obtained by the original simple random sample estima-
ApproShapley-opt estimation method. We start by determining the
tion approach ApproShapley and expression (5). 
sample size m that guarantees an estimation error of Shi lower
sim
Proposition 3.3. Let mn be the sample size used in the ApproShap- than e with a probability greater than 1 − α for all i ∈ N. We
ley stage and let mst , i = 1, . . . , n  = 1, . . . n, be the sample sizes end by comparing our results with those obtained by Maleki et al.
i
used in the second stage of the efficient stratified approach described (2014) using their proposed stratified estimation method.
 st,e f f First, we show some cases where the process gives the exact
above for estimating Shi . Then, the estimator Shi is unbiased, i.e.,
value Shi .
st,e f f
E[Shi ] = Shi .
Proposition 4.1. If the marginal contributions of a player i ∈ N de-
st,opt
Proof 3.3. The proof is straightforward given the properties of the pend only on the position in which he arrives, then Shi = Shi .
sample mean as an estimator of the population mean. Note that in
the two-stage procedure described above, the samples are drawn Proof 4.1. The proof is straightforward given that the value of x(O)i

independently in each stage, and thus is constant in each stratum Pi . Then, Shi = Shi , and thus
 
 ˆ i] 
n
1  1 
n
E[Sh Shst,opt = Shi = Sh = Shi .
E[Shst,e ff
] = E[Shst,opt ] + v (N ) − E Shst,opt i n n i
i i
i∈N
i v (N ) =1 =1
  
 Shi
= Shi + v (N ) − Shi Obviously, if player i ∈ N is a dummy player (i.e., v(S ∪ i ) =
v (N )
i∈N v(S ) + v(i )), then the condition given above holds. It also holds
Shi if the game is symmetric, and then it holds for every player i ∈
= Shi + (v(N ) − v(N )) = Shi ,
v (N ) N. In addition, this condition is fulfilled in some non-trivial cases.
For example, if we consider the weighted voting (N, v) game with
for all i ∈ N.  quota q = 10 and four players with weights 2,3,5, and 4, respec-
tively, then v(S ) = 1 if i ∈ S wi ≥ 10, and v(S ) = 0 otherwise. In
At this point, in the following remark, we recall the method
this case, the marginal contributions of player 3 are x(O )3 = 0
proposed by Maleki et al. (2014) as a particular case of St-
ApproShapley.
1
Malaki et al. did not specify how the total sample size m is allocated among the
Remark 3.1 (Maleki et al.’s method). Maleki et al. (2014) pro- estimation of the n Shapley values Sh1 , . . . , Shn . Their reported values mki (denoted
posed basing the estimation of the Shapley value for player i on as mk in Maleki et al., 2014) did not depend on player i, so we assume that mi = m n
184 J. Castro et al. / Computers and Operations Research 82 (2017) 180–188

for every marginal contribution in P31 ∪ P32 and x(O )3 = 1 for every Proposition 4.3. Let ei > 0 be the maximum estimation error for
marginal contribution in P33 ∪ P34 . player i’s Shapley value Shi . Then,
In the sequel, we will consider that the conditions established  
in Proposition 4.1 do not hold and that we use an optimal alloca-
 2e2i
P | Shst,opt − Shi | ≤ ei ≥ 1 − 2exp −  . (11)
i n−1 (xmax −xmin )2
i i
tion, i.e.,
=2 n2 m i
σ2
mik = n ik m, ∀ k = 1, . . . , n, ∀ i = 1, . . . , n. (8) Proof 4.3. Note that
∈L j σ j
2
j=1 
n  x ( O )i
st,opt Shst,opt = .
Thus, assuming that for all i = 1, . . . , n is normally dis-
Shi i n · mi
=1 x(O )i ∈Mi
tributed, which holds for large samples according to the central
limit theorem, then we conclude with the following result regard- Thus, since Mi1 , . . . , Min are random samples, then each of its el-
ing the precision of the estimate. ement is bounded by the two values xi min
and xi
max , so inequality
(11) holds (see Hoeffding, 1963, Theorem 2). 
Proposition 4.2. Let 1 − α be a given confidence level and e > 0 is
the maximum estimation error for player i’s Shapley value Shi for all Remark 4.2. Maleki et al. (2014) used the estimation error (11) to
i = 1, . . . , n. If the sampling size m verifies: determine the sample sizes mi , ı = 1, . . . , n,  = 1, . . . , n, allocated
n  to each stratum in order to minimize the estimation error within
zα2 /2 max j=1,...,n |L j | j=1 ∈L j σ j2
m≥ , (9) each of them. In particular, they first derived an upper bound for
n2 · e2 the marginal contributions range xi i
max − xmin , for each  = 1, . . . , n,
where zα /2 is the value such that P {Z ≥ zα /2 } = α2 , and Z is the stan- and they then determined a global upper bound for the estimation
 error of Shi . They set the values m∗k
dard normal distribution, then P |Shi − Shi | ≤ e ≥ 1 − α for all in (6) to minimize the global
st,opt
i
i = 1, . . . , n. upper bound.
The main problem with their allocation m∗k i
is that it does not
Proof 4.2. The proof is a straightforward application of the central depend on either the specific player i ∈ N or the special features of
limit theorem, Proposition 3.1, and expression (8) above.  the game. Moreover, it increases with position k by ignoring that
In general, the true variances σ j 2 ,  = 2, . . . , n − 1 are not σin2 = 0 for every player and for every game. The root cause of this
problem is the upper bound with which they work because it is
known. Thus, the variances must be estimated in order to deter-
excessively coarse. For example, if we consider the airport game
mine the minimum sample size, m, required to guarantee a given
(N, c) described in Section 6 (page 6), the range of the marginal
error with a certain probability.2 However, following the same ar-
contributions within each stratum are obtained as follows. Let F0 =
gument stated by Castro et al. (2009), we can give the upper 
j j 0 and Fk = ki=1 fi , where fi is the number of planes for which the
bounds on the estimation error e. Let xmax and xmin denote the
cost is ci = i (i = 1, . . . , 10 in this example). Then, since fi ≥ 2 for all
greatest and least marginal contribution of player j in Pj , respec-
i, then it holds that xi min
= 0 for all  ≥ 2 and for all i ∈ {1, . . . , 10}.
tively. Then, the following inequality holds:
With respect to the maximal marginal contributions, the following
(xmax
j j
− xmin )2 relations hold for every 2 ≤ i ≤ 10:
σ j2 ≤ , ∀ = 2, . . . , n − 1, ∀ j = 1, . . . , n.
4 xi
max = i − k, for all Fk−1 + 2 ≤  ≤ Fk + 1,
Therefore, when the confidence level 1 − α is given and assuming for all k = 1, . . . , i − 1. (12)
the optimal allocation, the estimation error can be bounded by
 Thus, for every fixed i ∈ {2, . . . , 10}, the range of the marginal con-
n  tributions within each stratum decreases with , so a good alloca-
zα /2 max j=1,...,n |L j | j=1 ∈L j (xmax − xmin )
j j 2

e≤ √ . (10) tion of the mi samples should prescribe smaller sample sizes for
2n m the latest positions. In Section 6, we use this game, among others,
Remark 4.1. If we need to reduce the estimation error, we can in order to measure the improvement obtained due to the optimal
−1 allocation of the m samples compared with that obtained using the
also compare the amounts mi and n−1 , which is an upper bound
of the cardinality of the different marginal contributions in Pi . allocation proposed by Maleki et al.
−1
 
Let Ki := { ∈ {1, . . . , n} | mi < n−1 }. Thus, for all i = 1, . . . , n and 5. Algorithm description
 ∈ Ki , we can exactly calculate
 (  − 1 )! ( n −  )! As mentioned earlier, the true variances σi2 must be estimated
Shi = (v(S ∪ i ) − v(S )). to determine the optimum sample sizes for each stratum Pi . In
S⊆N\i
( n − 1 )!
|S|=−1 this section, we describe our proposed method for use in practice,
which is a two-stage approach that divides the (total) sample size
In this case, the bound above (10) can be restated as follows: m into two parts, m = mexp + mst . Then, mexp samples are used to
   obtain unbiased estimates of the true variances σi2 , and we also
zα /2 max j=1,...,n (n − k j ) nj=1 ∈K j (xmax
j j
− xmin )2
employ this sample to estimate the marginal contributions of all
e≤  n ,
2n m − j=1 m(K j ) players in all positions (the strata in the sampling model) using the
 n−1 same sample size. Next, the remaining mst are used in the strata
where k j = |K j | and m(K j ) = ∈K j −1 . where the optimal sample gives a greater value than that assigned
in the previous step.
All the previous bounds are based on the central limit theo- Obviously, the selection of mexp and mst should also be decided
rem. Following Maleki et al. (2014), we can derive an alternative according to the characteristics of the current problem. However,
non-asymptotic bound, which is based on Hoeffding’s inequality we consider that the effort required to select these sizes is gen-
(Hoeffding, 1963). erally not worthwhile. Thus, we employ a balanced percentage of
50% in each part of the analysis to obtain the computational results
2
These estimations are also necessary for determining the optimal allocations. In in Section 6. Moreover, other features of the two-stage method
Section 5, we present the estimation process that we apply in practice. must be set, as follows.
 J. Castro et al. / Computers and Operations Research 82 (2017) 180–188 185

• The sampling size in Stage 1 is the same in all strata Pi , i.e.,
exp
mi = 2m n2
.
• The total sampling size in each stratum is proportional to the
estimated variance s2i . The sampling size mst i
for each stratum
Pi ,  = 1, . . . , n, in the stratified paper is calculated as follows:

s2i
mi = m n  n 2
,
i=1 k=1 sik

exp
mst
i = mi − mi .

Note that the sample sizes for each stratum are calculated over
the total sample size m because we use the information about
the marginal contributions x(O)i obtained from the pilot study
to give the estimation of Shi . Thus, in the second stage, only
exp
the remaining mi − mi samples are selected randomly. In the
st
case where mi ≤ 0, no more samples are generated for the
0

stratum Pi 0 and the sample size that we have not yet used
(−msti
) is allocated among the remaining strata in a propor-
0
exp
tional manner, mi − mi . We iterate this process until all mst
i
≥
0.

Next, we describe the method proposed to estimate Shi , which

we call Two-Stage-St-ApproShapley-opt and we denote it simply as
st,opt
ˆ
Sh in the sequel. Let m be the total sample size used to esti-
mate Sh.

• Stage 1. For each position  ∈ {1, . . . , n} and for each player i ∈

{1, . . . , n}:
1. The population of the sampling process is Pi and each sam-
pling unit represents a marginal contribution x(O )i ∈ Pi .
2. The parameters under paper are σi2 and Sh.
3. The characteristic observed in each sampling unit is x(O )i ∈
Pi .
exp exp
4. We select a random sample Mi of size mi = 2m n2
, which
is obtained with replacement.
exp Fig. 1. Two-Stage-St-ApproShapley-opt pseudo-code.
5. Let Shi be the mean of the marginal contributions over the
exp
sample Mi and let s2i be the unbiased sample variance of
exp
the marginal contributions over the sample Mi , i.e., 6. Finally, the estimate of the parameter Sh obtained by strati-
st,opt
fied sampling is Shˆ for all i = 1, . . . , n, where
i
exp 1   
Shi = x ( O )i , ∀ = 1, . . . , n, ∀i = 1, . . . , n. 
n
mexp
exp
Shi + mst
st
Shi
mexp ˆ
st,opt 1 i i
i x (O ) exp
i ∈Mi Sh i = .
n mexp + mst
1   
exp 2 =1 i i
s2i = x ( O )i − Shi ,
mexp −1 The pseudo-code for the proposed algorithm is shown in Fig. 1.
i x(O )i ∈Mi
exp

∀ = 1, . . . , n, ∀i = 1, . . . , n. 6. Computational results

• Stage 2. For each position  ∈ {1, . . . , n} and for each player i ∈ We used the same examples employed by Castro et al.
{1, . . . , n}: (2009) as a benchmark in order to measure the improvement ob-
1. The population in the sampling process is Pi and each sam- tained due to stratification and to compare our allocation method
pling unit represents a marginal contribution x(O )i ∈ Pi . with that of Maleki et al. We selected five well-known games from
2. The parameter under paper is Sh. the literature where the worth can be calculated in polynomial
3. The characteristic observed in each sampling unit is x(O )i ∈ time for any coalition with known Shapley values. We found that
Pi . method proposed by Maleki et al., which is also based on stratifi-
4. We select a random sample Mi st of size mst , which is ob-
i cation, improved the performance compared with the original es-
tained with replacement from each stratum Pi . timation method described by Castro et al. (2009) based on sim-
st
5. Let Shi be the mean of the marginal contributions over the ple random sampling, but it performed poorly compared with the
st (if mst = 0 then Sh = 0), i.e., st two-stage optimal allocation method proposed in the present pa-
sample Mi i i
per. We explored the factors that contribute to the relevant gains
 based on these examples, which showed that the gains increased
st 1
Shi = x ( O )i , ∀ = 1, . . . , n, ∀i = 1, . . . , n. with the dependence of the marginal contributions on the order of
mst entrance by players and when players are asymmetric.
i x(O )i ∈Mi
st
186 J. Castro et al. / Computers and Operations Research 82 (2017) 180–188

First, we recall the examples used by Castro et al. (2009).

Table 1
Errors in the symmetric voting game.
1. A symmetric voting game. Let (N, v) be a voting game where
N = {1, . . . , 10 0 0} and ∀ S ⊆ N, m/n 103 104 105 106 107 108
 esim 4 1 0.34 0.106 0.036 0.011
1 if |S| > 500, max
v (S ) = esim
ave 0.684 0.252 0.083 0.026 0.008 0.003
0 otherwise. sim
eth 31 9.8 3.1 0.98 0.31 0.098
est
max 0 0 0 0 0 0
Given that the players in this game are symmetric, it is easy to 100% 100% 100% 100% 100% 100%
see that the Shapley value is 0.001 for each of them. est
ave 0 0 0 0 0 0
2. A non-symmetric voting game. Let (N, v) be the non- 100% 100% 100% 100% 100% 100%
st
eth 0 0 0 0 0 0
symmetric game defined by Owen (1995) for a voting process 100% 100% 100% 100% 100% 100%
in the United States. The player set is N = {1, . . . , 51} and the eMa
max 0 0 0 0 0 0
characteristic function in this game is given by 100% 100% 100% 100% 100% 100%
   eMa
ave 0 0 0 0 0 0
1, if i∈S wi > j∈N w j /2,
100% 100% 100% 100% 100% 100%
v (S ) =
0, otherwise,

where w = {w1 , . . . , w51 } = {45, 41, 27, 26, 26, 25, 21, 17, 17, 14,
Table 2
13, 13, 12,12, 12, 11, 10, . . . , 10,9, . . . , 9, 8, 8, 7, . . . , 7, 6, . . . , 6, 5,
            Errors in the non-symmetric voting game.
4 times 4 times 4 times 4 times m/n 103 104 105 106 107 108
4, . . . , 4, 3, . . . , 3}.
      esim
max 8.83 4.495 0.969 0.283 0.115 0.053
9 times 7 times esim 3.104 1.073 0.321 0.082 0.03 0.014
ave
3. An airport game. Let (N, c) be an airport game where N = sim
eth 31 9.8 3.1 0.98 0.31 0.098
{1, . . . , 100}. The characteristic function in this game, v, is de- est
max 8.468 2.301 0.737 0.183 0.0822 0.0239
fined as follows, ∀ S ⊆ N, 4.10% 48.80% 23.91% 35.42% 28.53% 54.87%
est
ave 2.608 0.836 0.223 0.0613 0.0208 0.00667
c (S ) = max ci , 15.97% 22.12% 30.57% 25.29% 30.6% 52.39%
i∈S st
eth 19.777 6.254 1.978 0.625 0.198 0.0625
36.19% 36.19 % 36.19% 36.19% 36.19% 36.19%
where c = {c1 , . . . , c100 } = {1, . . . , 1, 2, . . . , 2, 3, . . . , 3, 4, . . . , 4,
            eMa
max 13.572 2.814 1.315 0.439 0.110 0.0553
8 times 12 times 6 times 14 times −34.94% 37.40% −26.31% −35.54% 4.35% −4.16%
5, . . . , 5, 6, . . . , 6, 7, . . . , 7, 8, . . . , 8, 9, . . . , 9, 10, . . . , 10}. eMa 2.964 0.938 0.338 0.117 0.0286 0.0118
                 
ave
4.51% 12.58% 5.03% −29.91% 4.67% 15.71%
8 times 9 times 13 times 10 times 10 times 10 times
4. A shoes game. Let (N, v) be a shoes game where N =
{1, . . . , 100}. The characteristic function in this game is defined
as follows, ∀ S ⊆ N, Table 3
Errors in the airport game.
v(S ) = min{|Sle f t |, |Sright |}, m/n 103 104 105 106 107 108
where |Sleft | is the number of left shoes and |Sright | the num- esim
max 66.63 37.26 6.21 1.91 0.64 0.24
ber of right shoes. In this example, |Sle f t | = |Sright | = 50. In the esim
ave 14.41 5.35 1.21 0.48 0.15 0.05
sim
sampling process, without loss of generality, it is assumed that eth 310 98 31 9.8 3.1 0.98
est
max 47.370 3.450 0.920 0.430 0.0995 0.0353
the first 50 players correspond to the left shoes and the other 28.91% 90.74% 85.18% 77.48% 84.46% 85.30%
50 to the right shoes. It is easy to see that the Shapley value of est
ave 5.437 0.646 0.190 0.0570 0.0201 0.00617
1
this game is Shi = ∀i ∈ N. st
62.27% 87.93% 84.30% 88.13% 86.63% 87.77%
2 eth 83.041 26.260 8.304 2.626 0.831 0.262
5. A minimum spanning tree game. Let G = (N ∪ {0}, E ) be a val- 73.21% 73.21% 73.21% 73.21% 73.21% 73.21%
ued graph where N = {1, . . . , 100} and the cost associated with eMa
max 65.630 30.414 7.753 2.260 1.052 0.321
1.50% 18.37% −19.90% −15.49% −39.16% −25.23%
an edge (i, j) is
eMa
ave 11.340 4.458 1.315 0.493 0.147 0.0521
Ci j := 21.30% 16.67% −7.98% −2.64% 2.00% −4.03%

1, if i = j + 1, i = j − 1, i = 1 ∧ j = 100, i=100 ∧ j=1,
101, if i = 0 or j = 0,
Table 4
∞, otherwise.
Errors in the shoes game.
A minimum spanning tree game (N, c) is a cost game, m/n 103 104 105 106 107 108
where for a given S ⊆ N, c(S) is the sum of the
esim 46 12.85 2.85 1.265 0.355 0.1255
edge cost of the minimum spanning tree, i.e., c (S ) = max
esim 10.95 3.85 0.95 0.31 0.115 0.035
ave
minimum spanning tree of the graph G|S∪{0} , which is the par- sim
eth 310 98 31 9.8 3.1 0.98
tial graph restricted to the players in coalition S and the source est
max 32.277 12.061 3.653 1.266 0.336 0.103
node 0. It is easy to see that the Shapley value for this game is 29.83% 5.92% −21.98% −0.08% 5.35% 18.19%
est
Shi = 2, ∀i ∈ N. ave 10.632 4.237 1.218 0.380 0.100 0.0356
2.90% −9.13% −22.00% 18.42% 13.04% −1.44%
st
eth 30.371 9.604 3.037 0.960 0.304 0.0960
We compared the estimations obtained by stratified random 2.00% 2.00% 2.00% 2.00% 2.00% 2.00%
sampling with the optimum allocation, the results obtained us- eMa
max 44.612 18.812 4.525 1.620 0.121 0.549
ing the original algorithm based on simple random sampling, and 3.02% −31.69% −37.02% −21.91% −35.34% 3.59%
those obtained by employing the allocation method of Maleki eMa
ave 12.777 4.813 1.505 0.546 0.155 0.0390
−14.30% −20.01% −36.88% −43.22% −25.81% −10.26%
et al. Tables 1–5 show the maximum error, average error, and theo-
retical error bound for these estimates, those based on simple ran-
 J. Castro et al. / Computers and Operations Research 82 (2017) 180–188 187

Table 5 Nࢨi can be –99, 1, or 101, depending on whether player i + 1

Errors in the minimal spanning tree game.
and i − 1 belong to S. Thus, for every position 3 ≤  ≤ 98,
m/n 103 104 105 106 107 108 xi i i2 i2
max = 101 and xmin = −99, whereas xmax = 101, xmin = 1 and
i 99 i 99
xmax = 1, xmin = −99. Thus,
esim
max 70 0 0 1700 625 202.4 79 21.4
esim
ave 2454 598.2 186.4 62.4 24 6.6 
sim
eth 6230 1970 623 197 62.3 19.7
st
z α2 100 · 98(2 · 1002 + 96 · 2002 )
est
max 5486.68 1312.75 472.47 168.10 46.37 16.33 eth = √ .
21.62% 22.78% 24.40% 16.94% 41.31% 23.69% 2 · 100 m
est
ave 1389.93 431.27 150.79 50.24 12.84 3.77
43.36% 27.91% 19.10% 19.48% 46.49% 42.92% The computational experiments yielded the expected results.
st
eth 6027.44 1906.04 602.74 190.60 60.27 19.06 From a general viewpoint, the figures in Tables 1–5 demonstrate
3.24% 3.24% 3.24% 3.24% 3.24% 3.24% the strong relationship between the gain due to stratification and
eMa 6421.70 1533.70 357.89 168.31 44.07 13.90
max
the dependence of the variability of the marginal contributions of
8.26% 9.78% 42.74% 16.84% 44.22% 35.05%
eMa 1636.66 449.18 150.66 44.25 15.03 3.79 each player on their entrance order.
ave
33.31% 24.91% 19.17% 29.09% 37.38% 42.58% In particular, we can see that the original symmetric algorithm
based on simple random sampling performs efficiently when the
following two conditions are generally satisfied, i.e., the variabili-
dom sampling3 (esim sim sim ties of the marginal contributions of each player are similar and
max , eave , and eth ), and the proposed improve-
ment based on stratified random sampling with the optimal allo- the marginal contributions are not highly dependent on the po-
cation (est st st sition in which player arrives. Thus, for any player i ∈ N, their
max , eave and eth ), as well as using the allocation of Maleki
marginal contribution to the coalition S ⊆ Nࢨ{i} (i.e., v(S ∪ {i} ) −
et al. (emax and eave ). The figures are given in 10−3 units. (Castro
Ma Ma
v(S )) does not depend strongly on the coalition S size, as found
et al., 2009) used α = 0.01 to calculate the theoretical error bound
with the shoes example.
(eth ). We note that this error represents the value that satisfies:
Moreover, based on the computational experiments and the
P (|Shi − Shˆ i | ≤ e ) ≥ 1 − α = 0.99. To calculate the new theoretical
th
st , we considered each of the five examples in the follow- construction of the algorithm, we can deduce that the algorithm
errors eth
proposed by Maleki et al. (2014) performs well when the vari-
ing figures. The theoretical error displayed in the tables was ob-
abilities of the marginal contributions of each player are similar
tained using (10).
(which is also the case for the symmetric algorithm). In addition,
1. Symmetric voting game: xi i
max = xmin , for all , and for all i ∈
and given the manner in which the sample is allocated among
N ⇒ ethst = 0. the strata in the algorithm of Maleki et al., the variability of each
2. Non-symmetric voting game: player’s marginal contributions should increase with the size of
∈ {45, 41}, xi max = xmin , for all  ∈ {1, . . . , 10} ∪
• w i the coalition so the last positions are more significant in order to
i
{40, . . . , 51}; and ximax − ximin = 1, otherwise. obtain more efficient performance by the algorithm. This was the
case for the minimal spanning tree game example.
max = xmin , for all  ∈ {1, . . . , 10} ∪
• w ∈ {27, 26, 25, 21, 17}, xi i
i
In summary, according to this comparative analysis, our pro-
{43, . . . , 51}; and xmax − xmin = 1, otherwise.
i i
posed algorithm significantly improves the performance compared
• w ∈ {10, . . . , 14}, xi i for all  ∈ {1, . . . , 10} ∪
i max = xmin , with the two previous algorithms in situations where the variabil-
{44, . . . , 51}; and xmax − xmin = 1, otherwise. Then,
i i
ity of the marginal contributions of each player are very different,
• w ∈ {3, . . . , 9}, xi i  ∈ {1, . . . , 11} ∪
i max = xmin , for all or in situations where the variability of the marginal contributions
{44, . . . , 51}; and xmax − xmin = 1, otherwise.
i i
depends greatly on the position in which players arrive. We ob-
Then, served these situations in some of the examples analyzed in this
 paper, such as the airport and non-symmetric voting games, which
st
z α2 33(2 · 29 + 7 · 32 + 11 · 33 + 31 · 32 ) showed that the proposed method obtained clear improvements
eth = √ .
2 · 51 m compared with the other methods. It was remarkable that the pro-
posed algorithm also performed very well (and was not worse than
3. Airport game: According to expression (12) of xi max , |Li | = Fi−1 ,
the other two) in the situations where the symmetric algorithm
for each i = 1, . . . , n, and thus it is verified that :
 and algorithm of Maleki et al. also performed well because we
10 i−1 made no assumption regarding the variability of the marginal con-
z α2 maxi=1,...,n Fi−1 i=1 fi k=1 f k ( i − k )2
st
eth = √ tributions used to estimate the Shapley value.
2 · 100 m For the symmetric voting game (see Table 1), considering
√
z α2 7180110 Proposition 4.1, both of the stratified methods yielded the exact
= √ . value Sh. Clearly, both of the stratified methods perform equally
2 · 100 m
well and they are better than the symmetric method based on
4. Shoes game: xi i
max − xmin = 1, for all 2 ≤  ≤ 99, and for all i ∈ simple random sampling. For the non-symmetric voting game (see
N. Then, Table 2), the improvement with our method ranged from 4.10%
√ to 54.87%, and a third of the original error was reduced on av-
z α2 100 · 982
st
eth = √ . erage. The method of Maleki et al. gave similar results to the orig-
2 · 100 m inal method and it was clearly outperformed by our proposed new
5. Minimum spanning tree game: For every player i ∈ N = method. The best results were obtained in the airport game (see
{1, . . . , 100}, their marginal contribution to the coalition S ⊆ Table 3), where the order strongly affected the marginal contribu-
tions of each player. The original error was around eight times the
new error and the improvement was generally greater than 80%.
3
These figures were obtained from Castro et al. (2009). We note that there are Again, the method of Maleki et al. gave similar results to the origi-
errors in Tables 1 and 3 in that paper. The theoretical errors presented for the nal, but it was clearly outperformed by our proposed new method.
symmetric and non-symmetric voting games in these tables were incorrect. In the
present paper, we provide the correct theoretical errors for these examples. How-
Table 4 shows that in this case, the improvement was negligible
ever, these figures did not affect the conclusions of the original paper by Castro because the entrance order had a weak effect on the marginal con-
et al. (2009). tributions. We also note that the method of Maleki et al. obtained
188 J. Castro et al. / Computers and Operations Research 82 (2017) 180–188
the worst performance in this case. For the last game, as men-
tioned earlier, the method of Maleki et al. performed in a simi-
lar manner to our method, although both obtained improvements
compared with the original method based on simple random sam-
pling (see Table 5).
7. Conclusions
Cooperative game theory is a very important topic and it has
applications in many disciplines, such as economics, politics, bi-
ology, sociology, operations research, and social network analysis.
One of the main topics in cooperative game theory is the allocation
of the worth of the grand coalition among the players involved in
the game, where the Shapley value is the solution concept that is
used most often for this purpose. In this sense, the Shapley value is
useful when it is necessary to allocate the value that a set of play-
ers can achieve if they agree to cooperate. Nevertheless, calculating
the Shapley value has been an open topic in game theory, where
the complexity of its calculation is generally exponential, although
its calculus has been solved well in some particular cases. This is
one of the reasons why the Shapley value has not been applied
extensively to real and large-scale problems.
As mentioned in the Introduction, the problem of computing
Shapley value’s has been studied from two alternative perspectives
in recent years: algorithms to compute the Shapley value exactly
(e.g., see Bilbao et al., 20 0 0; Granot et al., 2002) and algorithms
that focus on its approximation, where the main aim is to “esti-
mate” the Shapley value within a reasonable computational time.
The second approach has been investigated significantly less often
than the first. However, considering the widespread application of
game theory in real-world problems, where exact solutions are of-
ten not possible, it is necessary to develop algorithms that can fa-
cilitate this approximation. It should be noted (as mentioned by
Maleki et al., 2014) that the first attempt to estimate the Shap-
ley value for a large class of games was the algorithm proposed
by Castro et al. (2009). Other attempts to estimate the Shapley
value were mostly for a specific class of problems. Thus, for a real
situation and real cooperative game, the Shapley value must be
estimated using a sampling approach. For example, Fatima et al.
(2006) proposed a randomized polynomial method for approxi-
mating the Shapley value of weighted voting and k-majority voting
games, while Crama and Leruth (2007) used a Monte Carlo simula-
tion to estimate the Shapley value for a complex voting game. The
second well-known approach for estimating the Shapley value for
a large class of games was proposed by Maleki et al. (2014), who
proposed the use of stratified random sampling to reduce the error
estimation in the original method described by Castro et al. They
also considered the same strata treated in the present paper.
In order to compare the three algorithms for Shapley value es-
timation in general games, the symmetric algorithm defined by
Castro et al. (2009) and the algorithm proposed by Maleki et al.
(2014) require (to obtain better results) that the variabilities of
the marginal contributions of all players are similar. Moreover, the
symmetric algorithm requires that the marginal contributions do
not depend greatly on the positions in which the players arrive.
The algorithm proposed by Maleki et al. also requires that the vari-
ability of the marginal contributions increases with the size of the
coalitions. Our proposed algorithm clearly obtains improvements
in situations where the variabilities in the marginal contributions
of each player are very different, or in those situations where the
variability of the marginal contributions depends greatly on each
player’s arrival position (which does not necessarily increase with
the size of coalitions, as found with the algorithm proposed by
Maleki et al.). In situations where the symmetric algorithms or that
proposed by Maleki et al. perform very well, our algorithm does
not obtain significantly worse results because we make no assump-
tions regarding the variability of the marginal contributions.
In this paper, we focused on the Shapley value. It is important
to note that the method used to estimate the Shapley value can be
extended in a very similar manner to estimate any semivalue.
Finally, in conclusion, we not that the proposed procedure can
be computed in a parallel manner because the random estima-
tion process in each stratum does not require information from
the other strata. It would not be difficult to obtain a pseudo-code
that works in parallel in order to reduce the complexity in a linear
manner.
Acknowledgement
This paper was partially supported by the Plan Nacional de I+D+i
of the Spanish Government under projects MTM2011-27892 and
MTM2015-70550.
References
Bilbao, JM, Fernández, JR, Jiménez Losada, A, López, JJ, 20 0 0. Generating functions
for computing power indices efficiently. TOP 8 (2), 191–213.
Bolus, S, 2011. Power indices of simple games and vector-weighted majority games
by means of binary decision diagrams. Eur. J. Oper. Res. 210 (2), 258–272.
Castro, J, Gómez, D, Tejada, J, 2008. A polynomial rule for the problem of sharing
delay costs in PERT networks. Comput. Oper. Res. 35 (7), 2376–2387.
Castro, J, Gómez, D, Tejada, J, 2009. Polynomial calculation of the Shapley value
based on sampling. Comput. Oper. Res. 36 (5), 1726–1730.
Chalkiadakis, G, Elkind, E, Wooldridge, M, 2012. Computational aspects of coop-
erative game theory. Synthesis Lectures on Artificial Intelligence and Machine
Learning. Morgan & Claypool.
Cochran, WG, 1977. Sampling Techniques. Wiley.
Conitzer, V, Sandholm, T, 2006. Complexity of constructing solutions in the core
based on synergies among coalitions. Artif. Intel. J. 170, 607–619.
Crama, Y, Leruth, L, 2007. Control and voting power in corporate networks: concepts
and computational aspects. Eur. J. Oper. Res. 178, 879–893.
Deng, X, Papadimitriou, CH, 1994. On the complexity of cooperative solution con-
cepts. Math.Oper. Res. 19 (2), 257–266.
Faigle, U, Kern, W, 1992. The Shapley value for cooperative games under precedence
constraints. Int. J. Game Theory 21 (3), 249–266.
Fatima, SS, Wooldridge, M, Jenniggs, NR, 2006. An analysis of the Shapley value and
its uncertainty for the voting game. Lect.NotesArtif.Intell. 3937, 85–98.
Fernández, JR, Algaba, E, Bilbao, JM, Jiménez, A, Jiménez, N, López, JJ, 2002. Gener-
ating functions for computing the Myerson value. Ann Oper Res 109, 143–158.
Granot, D, Kuipers, J, Chopra, S, 2002. Cost allocation for a tree network with het-
erogeneous customers. Math. Oper. Res. 27 (4), 647–661.
Hoeffding, W, 1963. Probability inequalities for sums of bounded random variables.
J. Am. Stat. Assoc. 58 (301), 13–30.
Lohr, H, 1999. Sampling: Design and Analysis. Duxbury.
Lucchetti, R, Moretti, S, Patrone, F, Radrizzani, P, 2010. The Shapley and Banzhaf
values in microarray games. Comput. Oper. Res. 37, 1406–1412.
Maleki, S., Tran-Thanh, L., Hines, G., Rahwan, T., Rogers, A., 2014. Bounding the esti-
mation error of sampling-based Shapley value approximation arXiv:1306.4265.
Mann, I, Shapley, LS, 1960. Values for Large Games IV: Evaluating the Electoral Col-
lege by Monte Carlo Techniques. Technical report. The RAND Corporation, Santa
Monica.
Moretti, S, 2010. Statistical analysis of the Shapley value for microarray games. Com-
put. Oper. Res. 37, 1413–1418.
Neyman, J, 1934. On the two different aspects of the representative method: the
method of stratified sampling and the method of purposive selection. J. R. Stat.
Soc. 97, 558–606.
Owen, G, 1995. Game Theory. Academic press.
Shapley, LS, 1953. A value for n-person games. In: Kuhn, HW, Tucker, AW (Eds.),
Contributions to the Theory of Games II (Annals of Mathematics Studies 28).
Princeton University Press, pp. 307–317.
Weber, RJ, 1988. Probabilistic values for games. In: Roth, A (Ed.), The Shapley Value:
Essays in Honor of Lloyd S. Shapley. Cambridge University Press, pp. 101–119.