JAMDSJournal of Applied Mathematics and Decision Sciences1532-76121173-9126Hindawi Publishing Corporation34208910.1155/2009/342089342089Research ArticleConvex Interval GamesAlparslan GökS. Z.1, 2BranzeiR.3TijsS.4, 5WoodGraham1Institute of Applied MathematicsMiddle East Technical University06531 AnkaraTurkeymetu.edu.tr2Department of MathematicsFaculty of Arts and SciencesSüleyman Demirel University 32260 IspartaTurkeysdu.edu.tr3Faculty of Computer ScienceAlexandru Ioan Cuza University700483 IaşiRomaniauaic.ro4CentER and Department of Econometrics and ORTilburg UniversityP.O. Box 90153 5000LE TilburgThe Netherlandstilburguniversity.nl5Department of MathematicsUniversity of Genoa16126 GenoaItalyunige.it200923062009200923102008240320092009Copyright © 2009This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Convex interval games are introduced and characterizations are given. Some economic situations leading to convex interval games are discussed. The Weber set and the Shapley value are defined for a suitable class of interval games and their relations with the interval core for convex interval games are established. The notion of population monotonic interval allocation scheme (pmias) in the interval setting is introduced and it is proved that each element of the Weber set of a convex interval game is extendable to such a pmias. A square operator is introduced which allows us to obtain interval solutions starting from the corresponding classical cooperative game theory solutions. It turns out that on the class of convex interval games the square Weber set coincides with the interval core.

1. Introduction

In classical cooperative game theory payoffs to coalitions of players are known with certainty. A classical cooperative game is a pair N,v where N={1,2,,n} is a set of players and v:2N is a map, assigning to each coalition S2N a real number, such that v(Ø)=0. Often, we also refer to such a game as a (transferable utility) TU game. We denote by GN the family of all classical cooperative games with player set N. The class of convex games  is one of the most interesting classes of cooperative games from a theoretical point of view as well as regarding its applications in real-life situations. A game vGN is convex (or supermodular) if and only if the supermodularity condition v(ST)+v(ST)v(S)+v(T) for each S,T2N holds true. Many characterizations of classical convex games are available in literature (, Biswas et al. , Brânzei et al. , Martinez-Legaz [5, 6]). On the class CGN of classical convex games solution concepts have nice properties; for details we refer the reader to Brânzei et al. . Classical convex games have many applications in economic and real-life situations. It is well-known that classical public good situations , sequencing situations (Curiel et al. ), and bankruptcy situations (, Aumann and Maschler , Curiel et al. ) lead to convex games.

However, there are many real-life situations in which people or businesses are uncertain about their coalition payoffs. Situations with uncertain payoffs in which the agents cannot await the realizations of their coalition payoffs cannot be modelled according to classical game theory. Several models that are useful to handle uncertain payoffs exist in the game theory literature. We refer here to chance-constrained games (Charnes and Granot ), cooperative games with stochastic payoffs (Suijs et al. ), cooperative games with random payoffs (Timmer et al. . In all these models probability and stochastic theory plays an important role.

This paper deals with a model of cooperative games where only bounds for payoffs of coalitions are known with certainty. Such games are called cooperative interval games. Formally, a cooperative interval game in coalitional form (Alparslan Gök et al. ) is an ordered pair N,w where N={1,2,,n} is the set of players, and w:2NI() is the characteristic function such that w(Ø)=[0,0], where I() is the set of all nonempty, compact intervals in . For each S2N, the worth set (or worth interval) w(S) of the coalition S in the interval game N,w is of the form [w̲(S),w¯(S)]. We denote by IGN the family of all interval games with player set N. Note that if all the worth intervals are degenerate intervals, that is, w̲(S)=w¯(S) for each S2N, then the interval game N,w corresponds in a natural way to the classical cooperative game N,v where v(S)=w̲(S) for all S2N.

Cooperative interval games are very suitable to describe real-life situations in which people or firms that consider cooperation have to sign a contract when they cannot pin down the attainable coalition payoffs, knowing with certainty only their lower and upper bounds. Such contracts should specify how the interval uncertainty with regard to the coalition values will be incorporated in the allocation of the worth w(N) before its uncertainty is resolved, and how the realization of the payoff for the grand coalition Rw(N) will be finally distributed among the players. Interval solution concepts for cooperative interval games are a useful tool to settle cooperation within the grand coalition via such (binding) contracts.

An interval solution concept on IGN is a map assigning to each interval game wIGN a set of n-dimensional vectors whose components belong to I(). We denote by I()N the set of all such interval payoff vectors. An interval allocation obtained by interval solution concept commonly chosen by the players before the interval uncertainty with regard to the coalition values is removed offers at this ex-ante stage an estimation of what individual players may receive, between two bounds, when the uncertainty on the reward of the grand coalition is removed in the ex post stage. We notice that the agreement on a particular interval allocation (I1,,In) based on an interval solution concept merely says that the payoff xi that player i will receive in the interim or ex post stage is in the interval Ii. This is a very weak contract to settle cooperation. Therefore, writing down in the contact how to transform the interval allocation of w(N) into an allocation (x1,,xn) in n of the realization R of w(N), iNxi=R, in a consistent way with (I1,,In), that is, I̲ixiI¯i for each iN, is mandatory (see Brânzei et al. ).

In this paper, we introduce the class of convex interval games and extend classical results regarding characterizations of convex games and properties of solution concepts to the interval setting. Some classical TU-games associated with an interval game wIGN will play a key role, namely, the border games N,w̲, N,w¯ and the length game N,|w|, where |w|(S)=w¯(S)-w̲(S) for each S2N. Note that w¯=w̲+|w|.

The paper is organized as follows. In Section 2 we recall basic notions and facts from the theory of cooperative interval games. In Section 3 we introduce supermodular and convex interval games and give basic characterizations of convex interval games. In Section 4 we introduce for size monotonic interval games the notions of interval marginal operators, the interval Shapley value and the interval Weber set and study their properties for convex interval games. Moreover, we introduce the notion of population monotonic interval allocation scheme (pmias) and prove that each element of the interval Weber set of a convex interval game is extendable to such a pmias. In Section 5 we introduce the square operator and describe some interval solutions for interval games that have close relations with existing solutions from the classical cooperative game theory. It turns out that on the class of convex interval games the interval core and the square interval Weber set coincide. Finally, in Section 6 we conclude with some remarks on further research.

2. Preliminaries on Interval Calculus and Interval Games

In this section some preliminaries from interval calculus and some useful results from the theory of cooperative interval games are given (Alparslan Gök et al. ).

Let I,JI() with I=[I̲,I¯], J=[J̲,J¯], |I|=I¯-I̲ and α+. Then,

I+J=[I̲,I¯]+[J̲,J¯]=[I̲+J̲,I¯+J¯];

αI=α[I̲,I¯]=[αI̲,αI¯].

By (i) and (ii) we see that I() has a cone structure.

In this paper we also need a partial substraction operator. We define I-J, only if |I||J|, by I-J=[I̲,I¯]-[J̲,J¯]=[I̲-J̲,I¯-J¯]. Note that I̲-J̲I¯-J¯. We recall that I is weakly better than J, which we denote by IJ, if and only if I̲J̲ and I¯J¯. We also use the reverse notation IJ, if and only if I̲J̲ and I¯J¯. We say that I is better than J, which we denote by IJ, if and only if IJ and IJ.

For w1,w2IGN we say that w1w2 if w1(S)w2(S), for each S2N. For w1,w2IGN and λ+ we define N,w1+w2 and N,λw by (w1+w2)(S)=w1(S)+w2(S) and (λw)(S)=λ·w(S) for each S2N. So, we conclude that IGN endowed with is a partially ordered set and has a cone structure with respect to addition and multiplication with nonnegative scalars described above. For w1,w2IGN with |  w1(S)||w2(S)| for each S2N, N,w1-w2 is defined by (w1-w2)(S)=w1(S)-w2(S).

Now, we recall that the interval imputation set (w) of the interval game w, is defined by (w)={(I1,,In)I()NiNIi=w(N),Iiw(i),iN}, and the interval core 𝒞(w) of the interval game w, is defined by 𝒞(w)={(I1,,In)(w)iSIiw(S),  S2N{Ø}}.

Here, iNIi=w(N) is the efficiency condition and iSIiw(S), S2N{Ø}, are the stability conditions of the interval payoff vectors.

A game wIGN is called -balanced if for each balanced map λ:2N{Ø}+ we have S2N{Ø}λ(S)w(S)w(N). We recall that a map λ:2N{Ø}+ is called a balanced map  if S2N{Ø}λ(S)eS=eN. Here, eN=(1,,1), and for each S2N, (eS)i=1 if iS and (eS)i=0 otherwise. It is easy to prove that if N,w is -balanced then the border games N,w̲ and N,w¯ are balanced. A game wIGN is -balanced if and only if 𝒞(w)Ø (in Alparslan Gök et al. [17, Theorem 3.1]). We denote by BIGN the class of -balanced interval games with player set N.

Let wIGN, I=(I1,,In),J=(J1,,Jn)(w) and S2N{Ø}. We say that I dominates J via coalition S, denoted by I  doms  J, if

IiJi for all iS;

iSIiw(S).

For S2N{Ø} we denote by D(S) the set of those elements of (w) which are dominated via S. I is called undominated if there does not exist J and a coalition S such that J  doms  I. The interval dominance core 𝒟𝒞(w) of wIGN consists of all undominated elements in (w). For wIGN a subset A of (w) is an interval stable set if the following conditions hold.

(Internal stability ) there does not exist I,JA such that I dom J or J dom I.

(External stability ) for each IA there exist JA such that J dom I.

It holds 𝒞(w)𝒟𝒞(w)A for all wIGN and A a stable set of w.

3. Supermodular and Convex Interval Games

We say that a game N,w is supermodular if w(S)+w(T)w(ST)+w(ST)S,T2N.

From formula (3.1) it follows that a game N,w is supermodular if and only if its border games N,w̲ and N,w¯ are supermodular (convex). We introduce the notion of convex interval game and denote by CIGN the class of convex interval games with player set N. We call a game wIGNconvex if N,w is supermodular and its length game N,|w| is also supermodular. We straightforwardly obtain characterizations of games wCIGN in terms of w̲, w¯ and |w|GN.

Proposition 3.1.

Let wIGN and its related games |w|,w̲,w¯GN. Then the following assertions hold.

A game N,w is convex if and only if its length game N,|w| and its border games N,w̲, N,w¯ are convex.

A game N,w is convex if and only if its border game N,w̲ and the game N,w¯-w̲ are convex.

We notice that the nonempty set CIGN is a subcone of IGN and traditional convex games can be embedded in a natural way in the class of convex interval games because if vGN is convex then the corresponding game wIGN which is defined by w(S)=[v(S),v(S)] for each S2N is also convex. The next example shows that a supermodular interval game is not necessarily convex.

Example 3.2.

Let N,w be the two-person interval game with w(Ø)=[0,0], w(1)=w(2)=[0,1] and w(1,2)=[3,4]. Here, N,w is supermodular, but |w|(1)+|w|(2)=2>1=|w|(1,2)+|w|(Ø). Hence, N,w is not convex.

The next example shows that an interval game whose length game is supermodular is not necessarily convex.

Example 3.3.

Let N,w be the three-person interval game with w(i)=[1,1] for each iN, w(N)=w(1,3)=w(1,2)=w(2,3)=[2,2], and w(Ø)=[0,0]. Here, N,w is not convex, but N,|w| is supermodular, since |w|(S)=0, for each S2N.

Interesting examples of convex interval games are unanimity interval games. First, we recall the definition of such games. Let JI() with J[0,0] and let T2N{Ø}. The unanimity interval game based on J and T is defined by uT,J(S)={J,TS,[0,0],otherwise, for each S2N.

Clearly, N,|uT,J| is supermodular. The supermodularity of N,uT,J can be checked by considering the following case study: uT,J(AB)uT,J(AB)uT,J(A)uT,J(B)TA,  TBJJJJTA,  TBJ[0,0]J[0,0]TA,  TBJ[0,0][0,0]JTA,  TBJor [0,0][0,0][0,0][0,0].

For convex TU-games various characterizations are known. In the next theorem we give some characterizations of convex interval games inspired by Shapley .

Theorem 3.4.

Let wIGN be such that |w|GN is supermodular. Then, the following three assertions are equivalent:

wIGN is convex;

for all S1,S2,U2N with S1S2NU one has w(S1U)-w(S1)w(S2U)-w(S2);

for all S1,S22N and iN such that S1S2N{i} one has w(S1{i})-w(S1)w(S2{i})-w(S2).

Proof.

We show (i)(ii), (ii)(iii), (iii)(i).

Suppose that (i) holds. To prove (ii) take S1,S2,U2N with S1S2NU. From (3.1) with S1U in the role of S and S2 in the role of T we obtain (3.4) by noting that ST=S2U, ST=S1. Hence, (i) implies (ii).

That (ii) implies (iii) is straightforward (take U={i}).

Now, suppose that (iii) holds. To prove (i) take S,T2N. Clearly, (3.1) holds if ST. In case TS, suppose that ST consists of the elements i1,,ik and let D=ST. Then, w(S)-w(ST)=w(D{i1})-w(D)+s=2k(w(D{i1,,is})-w(D{i1,,is-1}))w(T{i1})-w(T)+s=2k(w(T{i1,,is})-w(T{i1,,is-1}))=w(ST)-w(T),for each S2N,  where the inequality follows from (iii).

Next we give as a motivating example a situation with an economic flavour leading to a convex interval game.

Example 3.5.

Let N={1,2,,n} and let f:[0,n]I() be such that f(x)=[f1(x),f2(x)] for each x[0,n] and f(0)=[0,0]. Suppose that f1:[0,n], f2:[0,n] and (f2-f1):[0,n] are convex monotonic increasing functions. Then, we can construct a corresponding interval game w:2NI() such that w(S)=f(|S|)=[f1(|S|),f2(|S|)] for each S2N. It is easy to show that w is a convex interval game with the symmetry property w(S)=w(T) for each S,T2N with |S|=|T|.

We can see N,w as a production game if we interpret f(s) for sN as the interval reward which s players in N can produce by working together.

Before closing this section we indicate some other economic and OR situations related to supermodular and convex interval games. In case the parameters determining sequencing situations are not numbers but intervals, under certain conditions also convex interval games appear (Alparslan Gök et al. [17, 19]). Bankruptcy situations when the estate of the bankrupt firm and the claims are intervals, under restricting conditions, give rise in a natural way to supermodular interval games which are not necessarily convex .

4. The Shapley Value, the Weber Set and Population Monotonic Allocation Schemes

We call a game N,w  size monotonic if N,|w| is monotonic, that is, |w|(S)|w|(T) for all S,T2N with ST. For further use we denote by SMIGN the class of size monotonic interval games with player set N. We notice that size monotonic games may have an empty interval core. In this section we introduce interval marginal operators on the class of size monotonic interval games, define the Shapley value and the Weber set on this class of games, and study their properties on the class of convex interval games.

Denote by Π(N) the set of permutations σ:NN. Let wSMIGN. We introduce the notions of interval marginal operator corresponding to σ, denoted by mσ, and of interval marginal vector of w with respect to σ, denoted by mσ(w). The marginal vector mσ(w) corresponds to a situation, where the players enter a room one by one in the order σ(1), σ(2),,σ(n) and each player is given the marginal contribution he/she creates by entering. If we denote the set of predecessors of i in σ by Pσ(i)={rNσ-1(r)<σ-1(i)}, where σ-1(i) denotes the entrance number of player i, then mσ(k)σ(w)=w(Pσ(σ(k)){σ(k)})-w(Pσ(σ(k))), or miσ(w)=w(Pσ(i){i})-w(Pσ(i)). We notice that mσ(w) is an efficient interval payoff vector for each σΠ(N). For size monotonic games N,w, w(T)-w(S) is well defined for all S,T2N with ST since |w(T)|=|w|(T)|w|(S)=|w(S)|. Now, we notice that for each wSMIGN the interval marginal vectors mσ(w) are defined for each σΠ(N), because the monotonicity of |w| implies w¯(S{i})-w̲(S{i})w¯(S)-w̲(S), which can be rewritten as w¯(S{i})-w¯(S)w̲(S{i})-w̲(S). So, w(S{i})-w(S) is defined for each SN and iS.

The following example illustrates that for interval games which are not size monotonic it might happen that some interval marginal vectors do not exist.

Example 4.1.

Let N,w be the interval game with N={1,2}, w(1)=[1,3],w(2)=[0,0] and w(1,2)=[2,3(1/2)]. This game is not size monotonic. Note that m(12)(w) is not defined because w(1,2)-w(1) is undefined since |w(1,2)|<|w(1)|.

A characterization of convex interval games with the aid of interval marginal vectors is given in the following theorem.

Theorem 4.2.

Let wIGN. Then, the following assertions are equivalent:

w is convex;

|w| is supermodular and mσ(w)𝒞(w) for all σΠ(N).

Proof.

(i)(ii) Let wCIGN, let σΠ(N) and take mσ(w). Clearly, we have kNmkσ(w)=w(N). To prove that mσ(w)𝒞(w) we have to show that for S2N, kSmkσ(w)w(S). Let S={σ(i1),σ(i2),,σ(ik)} with i1<i2<<ik. Then, w(S)=w(σ(i1))-w(Ø)+r=2k(w(σ(i1),σ(i2),,σ(ir))-w(σ(i1),σ(i2),,σ(ir-1)))w(σ(1),,σ(i1))-w(σ(1),,σ(i1-1))+r=2k(w(σ(1),σ(2),,σ(ir))-w(σ(1),σ(2),,σ(ir-1)))=r=1kmσ(ir)σ(w)=kSmkσ(w), where the inequality follows from Theorem 3.4. Further, by convexity of w, |w| is supermodular.

(ii)(i) From mσ(w)𝒞(w) for all σΠ(N) follows that mσ(w̲)C(w̲) and mσ(w¯)C(w¯) for all σΠ(N). Now, by the well-known characterization of classical convex games with the aid of marginal vectors we obtain that N,w̲ and N,w¯ are convex games. Since N,|w| is convex by hypothesis, we obtain by Proposition 3.1(i) that N,w is convex.

Now, we straightforwardly extend for size monotonic interval games two important solution concepts in cooperative game theory which are based on marginal worth vectors: the Weber set  and the Shapley value .

The interval Weber set 𝒲 on the class of size monotonic interval games is defined by 𝒲(w)=conv{mσ(w)σΠ(N)} for each wSMIGN. We notice that for traditional TU-games we have W(v)Ø for all vGN, while for interval games it might happen that 𝒲(w)=Ø (in case none of the interval marginal vectors mσ(w) is defined). Clearly, 𝒲(w)Ø for all wSMIGN. Further, it is well-known that C(v)=W(v) if and only if vGN is convex. However, this result cannot be extended to convex interval games as we prove in the following proposition.

Proposition 4.3.

Let wCIGN. Then, 𝒲(w)𝒞(w).

Proof.

By Theorem 4.2 we have mσ(w)𝒞(w) for each σΠ(N). Now, we use the convexity of 𝒞(w).

The following example shows that the inclusion in Proposition 4.3 might be strict.

Example 4.4.

Let N={1,2} and let w:2NI() be defined by w(1)=w(2)=[0,1] and w(1,2)=[2,4]. This game is convex. Further, m(1,2)(w)=([0,1],[2,3]) and m(2,1)(w)=([2,3],[0,1]), belong to the interval core 𝒞(w) and 𝒲(w)=conv{m(1,2)(w),m(2,1)(w)}. Notice that ([1/2,1(3/4)],[1(1/2),2(1/4)])𝒞(w) and there is no α[0,1] such that αm(1,2)(w)+(1-α)m(2,1)(w)=([1/2,1(3/4)],[1(1/2),2(1/4)]). Hence, 𝒲(w)𝒞(w) and 𝒲(w)𝒞(w).

In Section 5 we introduce a new notion of Weber set and show that the equality between the interval core and that Weber set still holds on the class of convex interval games.

The interval Shapley value   Φ:SMIGNI()N is defined by Φ(w)=1n!σΠ(N)mσ(w),for each  wSMIGN.

Since Φ(w)𝒲(w) for each wSMIGN, by Proposition 4.3 we have Φ(w)𝒞(w) for each wCIGN. Without going into details we note here that the Shapley value Φ on the class of size monotonic interval games, and consequently on CIGN, satisfies the properties of additivity, efficiency, symmetry and dummy player, which are straightforward generalizations of the corresponding properties for classical TU-games.

Proposition 4.5.

Let wIGN. If N,w is convex, then it is size monotonic.

Proof.

Let wCIGN. This assures that N,|w| is supermodular which implies that N,|w| is monotonic because for each S,T2N with ST we have |w|(T)+|w|(Ø)|w|(S)+|w|(TS), and from this inequality follows |w|(S)|w|(T) since |w|(TS)0. So, CIGNSMIGN.

In the next two propositions we provide explicit expressions of the interval marginal vectors and of the interval Shapley value on SMIGN.

Proposition 4.6.

Let wSMIGN and let σΠ(N). Then, miσ(w)=[miσ(w̲),miσ(w¯)] for all iN.

Proof.

By definition, mσ(w̲)=(w̲(σ(1)),w̲(σ(1),σ(2))-w̲(σ(1)),,w̲(σ(1),,σ(n))-w̲(σ(1),,σ(n-1))),mσ(w¯)=(w¯(σ(1)),w¯(σ(1),σ(2))-w¯(σ(1)),,w¯(σ(1),,σ(n))-w¯(σ(1),,σ(n-1))). Now, we prove that mσ(w¯)-mσ(w̲)0. Since |w|=w¯-w̲ is a classical convex game we have for each kNmσ(k)σ(w¯)-mσ(k)σ(w̲)=(w¯-w̲)(σ(1),,σ(k))-(w¯-w̲)(σ(1),,σ(k-1))=|w|(σ(1),,σ(k))-|w|(σ(1),,σ(k-1))0, where the inequality follows from the monotonicity of |w|. So, miσ(w̲)miσ(w¯) for all iN, and ([miσ(w̲),miσ(w¯)])iN=(w(σ(1)),,w(σ(1),,σ(n))-w(σ(1),,σ(n-1)))=mσ(w).

Since CIGNSMIGN we obtain from Proposition 4.6 that miσ(w)=[miσ(w̲),miσ(w¯)] for each wCIGN, σΠ(N) and for all iN.

Proposition 4.7.

Let wSMIGN and let σΠ(N). Then, Φi(w)=[ϕi(w̲),ϕi(w¯)] for all iN.

Proof.

From (4.2) and Proposition 4.6 we have for all iNΦi(w)=1n!σΠ(N)miσ(w)=1n!σΠ(N)[miσ(w̲),miσ(w¯)]=[1n!σΠ(N)miσ(w̲),1n!σΠ(N)miσ(w¯)]=[ϕi(w̲),ϕi(w¯)]

From Proposition 4.7 we obtain that for each wCIGN we have Φi(w)=[ϕi(w̲),ϕi(w¯)] for all iN.

In the sequel we introduce the notion of population monotonic interval allocation scheme (pmias) for totally -balanced interval games, which is a direct extension of pmas for classical cooperative games . A game wIGN is called totally -balanced if the game itself and all its subgames are -balanced.

We say that for a game wTBIGN a scheme A=(AiS)iS,S2N{Ø} with AiSI()N is a pmias of w if

iSAiS=w(S) for all S2N{Ø};

AiSAiT for all S,T2N{Ø} with ST and for each iS.

Notice that the total -balancedness of an interval game is a necessary condition for the existence of a pmias for that game. A sufficient condition is the convexity of the interval game. We notice that all subgames of a convex interval game are also convex. In what follows we focus on pmias on the class of convex interval games.

We say that for a game wCIGN an imputation I=(I1,,In)(w) is pmias extendable if there exist a pmias A=(AiS)iS,S2N{Ø} such that AiN=Ii for each iN.

Theorem 4.8.

Let wCIGN. Then, each element I of 𝒲(w) is extendable to a pmias of w.

Proof.

Let wCIGN. First, we show that for each σΠ(N), mσ(w) is extendable to a pmias. We know that the interval marginal operator mσ:SMIGNI()N is efficient for each σΠ(N). Then, for each S2N, iSmiσ(wS)=kSmσ(k)σ(wS)=w(S) holds, where (S,wS) is the corresponding (convex) subgame.

Further, by convexity, miσ(wS)miσ(wT) for each iSTN, where (S,wS) and (T,wT) are the corresponding subgames.

Second, each I𝒲(w) is a convex combination of mσ(w), σΠ(N), that is, I=ασmσ(w) with ασ[0,1] and σΠ(N)ασ=1. Now, since all mσ(w) are pmias extandable, we obtain that I is pmias extendable as well.

From Theorem 4.8 we obtain that the total interval Shapley value, that is, the interval Shapley value applied to the game itself and all its subgames, generates a pmias for each convex interval game. We illustrate this in Example 4.9, where the calculations are based on Proposition 4.7.

Example 4.9.

Let wCIGN with w(Ø)=[0,0], w(1)=w(2)=w(3)=[0,0], w(1,2)=w(1,3)=w(2,3)=[2,4] and w(1,2,3)=[9,15]. It is easy to check that the interval Shapley value generates for this game the pmias depicted as N{1,2}{1,3}{2,3}{1}{2}{3}[123[3,5][3,5][3,5][1,2][1,2]*[1,2]*[1,2]*[1,2][1,2][0,0]***[0,0]***[0,0]].

5. Interval Solutions Obtained with the Square Operator

Let a=(a1,,an) and b=(b1,,bn) with ab. Then, we denote by ab the vector ([a1,b1],,[an,bn])I()N generated by the pair (a,b)N. Let A,BN. Then, we denote by AB the subset of I()N defined by AB={abaA,bB,ab}.

With the use of the operator, we give a procedure to extend classical multisolutions on GN to interval multisolutions on IGN.

For a multisolution :GNN we define :IGNI()N by =(w̲)(w¯) for each wIGN.

Now, we focus on this procedure for multisolutions such as the core and the Weber set on interval games. We define the square interval core 𝒞:IGNI()N by 𝒞(w)=C(w̲)C(w¯) for each wIGN. We notice that a necessary condition for the non-emptiness of the square interval core is the balancedness of the border games.

Proposition 5.1.

Let wBIGN. Then, 𝒞(w)=𝒞(w).

Proof.

(I1,,In)𝒞(w) if and only if (I̲1,,I̲n)C(w̲) and (I¯1,,I¯n)C(w¯) if and only if (I1,,In)=(I̲1,,I̲n)(I¯1,,I¯n)𝒞(w).

Since CIGNBIGN we obtain that 𝒞(w)=C(w̲)C(w¯) for each wCIGN.

We define the square Weber set 𝒲:IGNI()N by 𝒲(w)=W(w̲)W(w¯) for each wIGN. Note that 𝒞(w)=𝒲(w) if wCIGN.

The next two theorems are very interesting because they extend for interval games, with the square interval Weber set in the role of the Weber set, the well-known results in classical cooperative game theory that C(v)W(v) for each vGN  and C(v)=W(v) if and only if v is convex .

Theorem 5.2.

Let wIGN. Then, 𝒞(w)𝒲(w).

Proof.

If 𝒞(w)=Ø the inclusion holds true. Suppose 𝒞(w)Ø and let (I1,,In)𝒞(w). Then, by Proposition 5.4, (I̲1,,I̲n)C(w̲) and (I¯1,,I¯n)C(w¯), and, because C(v)W(v) for each vGN, we obtain (I̲1,,I̲n)W(w̲) and (I¯1,,I¯n)W(w¯). Hence, we obtain (I1,,In)𝒲(w).

From Theorem 5.2 and Proposition 4.3 we obtain that 𝒲(w)𝒲(w) for each wCIGN. This inclusion might be strict as Example 4.4 illustrates.

Theorem 5.3.

Let wBIGN. Then, the following assertions are equivalent:

w is convex;

|w| is supermodular and 𝒞(w)=𝒲(w).

Proof.

By Proposition 3.1(i), w is convex if and only if |w|,w̲ and w¯ are convex. Clearly, the convexity of |w| is equivalent with its supermodularity. Further, w̲ and w¯ are convex if and only if W(w̲)=C(w̲) and W(w¯)=C(w¯). These equalities are equivalent with 𝒲(w)=𝒞(w). By Proposition 5.1 this is equivalent to 𝒞(w)=𝒲(w).

With the aid of Theorem 5.3 we will show that the interval core is additive on the class of convex interval games, which is inspired by Dragan et al. .

Proposition 5.4.

The interval core 𝒞:CIGNI()N is an additive map.

Proof.

The interval core is a superadditive solution concept for all interval games (Alparslan Gök et al. ). We need to show the subadditivity of the interval core. We have to prove that 𝒞(w1+w2)𝒞(w1)+𝒞(w2). Note that mσ(w1+w2)=mσ(w1)+mσ(w2) for each w1,w2CIGN. By definition of the square interval Weber set we have 𝒲(w1+w2)=W(w̲1+w̲2)W(w¯1+w¯2). By Theorem 5.3 we obtain 𝒞(w1+w2)=𝒲(w1+w2)𝒲(w1)+𝒲(w2)=𝒞(w1)+𝒞(w2).

Finally, we define 𝒟𝒞(w)=DC(w̲)DC(w¯) for each wIGN and notice that for convex interval games we have 𝒟𝒞(w)=DC(w̲)DC(w¯)=C(w̲)C(w¯)=𝒞(w)=𝒞(w), where the second equality follows from the well-known result in the theory of TU-games that for convex games the core and the dominance core coincide, and the last equality follows from Proposition 5.1. From 𝒟𝒞(w)=𝒞(w) for each wCIGN and 𝒞(w)𝒟𝒞(w) for each wIGN we obtain 𝒟𝒞(w)𝒟𝒞(w) for each wCIGN. We notice that this inclusion might be strict (Alparslan Gök et al. , Example 4.1).

6. Concluding Remarks

In this paper we define and study convex interval games. We note that the combination of Theorems 3.4, 4.2 and 5.3 can be seen as an interval version in (Brânzei et al. , Theorem 96). In fact these theorems imply (Brânzei et al. , Theorem 96) for the embedded class of classical TU-games. Extensions to convex interval games of the characterizations of classical convex games where exactness of subgames and superadditivity of marginal (or remainder) games play a role (Biswas et al. , Brânzei et al.  and Martinez-Legaz [5, 6]) can be found in (Brânzei et al. ).

There are still many interesting open questions. For further research it is interesting to study whether one can extend to interval games the well-known result in the traditional cooperative game theory that the core of a convex game is the unique stable set . It is also interesting to find an axiomatization of the interval Shapley value on the class of convex interval games. An axiomatic characterization of the interval Shapley value on a special subclass of convex interval games can be found in Alparslan Gök et al. . Other topics for further research could be related to introducing new models in cooperative game theory by generalizing cooperative interval games. For example, the concepts and results on (convex) cooperative interval games could be extended to cooperative games in which the coalition values w(S) are ordered intervals of the form [u,v] of an (infinite dimensional) ordered vector space. Such generalization could give more applications to the interval game theory. Also to establish relations between convex interval games and convex games in other existing models of cooperative games could be interesting. One candidate for such study could be convex games in cooperative set game theory .

Acknowledgments

Alparslan Gök acknowledges the support of TUBITAK (Turkish Scientific and Technical Research Council) and hospitality of Department of Mathematics, University of Genoa, Italy. Financial support from the Government of Spain and FEDER under project MTM2008-06778-C02-01 is gratefully acknowledged by R. Branzei and S. Tijs. The authors thank Elena Yanovskaya for her valuable comments. The authors greatfully acknowledge two anonymous referees.

ShapleyL. S.Cores of convex gamesInternational Journal of Game Theory197111112610.1007/BF01753431MR0311338ZBL0222.90054DriessenT.Cooperative Games, Solutions and Applications1988Dordrecht, The NetherlandsKluwer Academic PublishersZBL0686.90043BiswasA. K.ParthasarathyT.PottersJ. A. M.VoorneveldM.Large cores and exactnessGames and Economic Behavior199928111210.1006/game.1998.0686MR1698988ZBL0957.91009BrânzeiR.DimitrovD.TijsS.Models in Cooperative Game Theory2008Berlin, GermanySpringerGame Theory and Mathematical MethodsZBL1142.91017Martinez-LegazJ. E.Two remarks on totally balanced games1997TR#317Arlington, Tex, USADepartment of Mathematics, The University of TexasMartínez-LegazJ. E.SeegerA.Some characterizations of convex gamesRecent Advances in Optimization2006563Berlin, GermanySpringer293303Lecture Notes in Economics and Mathematical SystemsMR219116410.1007/3-540-28258-0_18ZBL1152.91335MoulinH.Axioms of Cooperative Decision Making198815Cambridge, UKCambridge University Pressxiv+332Econometric Society MonographsMR1113263ZBL0699.90001CurielI.PederzoliG.TijsS.Sequencing gamesEuropean Journal of Operational Research198940334435110.1016/0377-2217(89)90427-XMR1004986ZBL0674.90107O'NeillB.A problem of rights arbitration from the TalmudMathematical Social Sciences19822434537110.1016/0165-4896(82)90029-4MR662176ZBL0489.90090AumannR. J.MaschlerM.Game theoretic analysis of a bankruptcy problem from the TalmudJournal of Economic Theory198536219521310.1016/0022-0531(85)90102-4MR804893ZBL0578.90100CurielI. J.MaschlerM.TijsS.Bankruptcy gamesZeitschrift für Operations Research. Serie A. Serie B1987315A143A159MR917932ZBL0636.90100CharnesA.GranotD.Prior solutions: extensions of convex nucleus solutions to chance-constrained gamesProceedings of the Computer Science and Statistics Seventh Symposium at Iowa State UniversityOctober 1973Ames, Iowa, USA323332SuijsJ.BormP.De WaegenaereA.TijsS.Cooperative games with stochastic payoffsEuropean Journal of Operational Research1999113119320510.1016/S0377-2217(97)00421-9ZBL0972.91012TimmerJ.BormP.TijsS.Convexity in stochastic cooperative situationsInternational Game Theory Review200571254210.1142/S0219198905000387MR2133091ZBL1105.91004Alparslan GökS. Z.BrânzeiR.TijsS.The interval Shapley value: an axiomatizationno. 130, Institute of Applied Mathematics, METU, Ankara, Turkey, 2009, to appear in Central European Journal of Operations Research (CEJOR)BrânzeiR.TijsS.Alparslan GökS. Z.How to handle interval solutions for cooperative interval gamespreprint no. 110, Institute of Applied Mathematics, METU, 2008Alparslan GökS. Z.BrânzeiR.TijsS.Cores and stable sets for interval-valued games20082008-17Tilburg, The NetherlandsCenter for Economic Research, Tilburg Universitypreprint no. 90TijsS.Introduction to Game Theory200323New Delhi, IndiaHindustan Book Agencyviii+176Texts and Readings in MathematicsMR1978948ZBL1018.91001Alparslan GökS. Z.BrânzeiR.FragnelliV.TijsS.Sequencing interval situations and related games20082008-63Tilburg, The NetherlandsCenter for Economic Research, Tilburg Universitypreprint no. 113BrânzeiR.Alparslan GökS. Z.Bankruptcy problems with interval uncertaintyEconomics Bulletin2008356110WeberR. J.RothA. E.Probabilistic values for gamesThe Shapley Value: Essays in Honour of Lloyd S. Shapley1988Cambridge, UKCambridge University Press101119MR989825ShapleyL. S.A value for n-person gamesContributions to the Theory of Games, Vol. 21953Princeton, NJ, USAPrinceton University Press307317Annals of Mathematics Studies, no. 28MR0053477ZBL0050.14404SprumontY.Population monotonic allocation schemes for cooperative games with transferable utilityGames and Economic Behavior19902437839410.1016/0899-8256(90)90006-GMR1082655ZBL0753.90083IchiishiT.Super-modularity: applications to convex games and to the greedy algorithm for LPJournal of Economic Theory198125228328610.1016/0022-0531(81)90007-7MR640200ZBL0478.90092DrăganI.PottersJ.TijsS.Superadditivity for solutions of coalitional gamesLibertas Mathematica19899101110MR1048250ZBL0684.90105BrânzeiR.DimitrovD.TijsS.A new characterization of convex games2004109Tilburg, The NetherlandsCenter for Economic Research, Tilburg UniversityBrânzeiR.TijsS.Alparslan GökS. Z.Some characterizations of convex interval gamesAUCO Czech Economic Review200823219226SunH.Contributions to set game theory, Ph.D. thesis2003Enschede, The NetherlandsUniversity of Twente, Twente University Press