Cooperative game

This article is about a part of game theory. For video gaming, see Cooperative gameplay. For the similar feature in some board games, see cooperative board game

In game theory, a cooperative game is a game where groups of players ("coalitions") may enforce cooperative behaviour, hence the game is a competition between coalitions of players, rather than between individual players. An example is a coordination game, when players choose the strategies by a consensus decision-making process.

Recreational games are rarely cooperative, because they usually lack mechanisms by which coalitions may enforce coordinated behaviour on the members of the coalition. Such mechanisms, however, are abundant in real life situations (e.g. contract law).

Mathematical definition

A cooperative game is given by specifying a value for every coalition. Formally, the game (coalitional game) consists of a finite set of players  N , called the grand coalition, and a characteristic function  v : 2^N \to \mathbb{R} [1] from the set of all possible coalitions of players to a set of payments that satisfies  v( \emptyset ) = 0 . The function describes how much collective payoff a set of players can gain by forming a coalition, and the game is sometimes called a value game or a profit game. The players are assumed to choose which coalitions to form, according to their estimate of the way the payment will be divided among coalition members.

Conversely, a cooperative game can also be defined with a characteristic cost function  c: 2^N \to \mathbb{R} satisfying  c( \emptyset ) = 0 . In this setting, players must accomplish some task, and the characteristic function  c represents the cost of a set of players accomplishing the task together. A game of this kind is known as a cost game. Although most cooperative game theory deals with profit games, all concepts can easily be translated to the cost setting.

Duality


Let  v be a profit game. The dual game of  v is the cost game  v^* defined as

 v^*(S) = v(N) - v( N \setminus S ), \forall~ S \subseteq N.\,

Intuitively, the dual game represents the opportunity cost for a coalition  S of not joining the grand coalition  N . A dual profit game  c^* can be defined identically for a cost game  c . A cooperative game and its dual are in some sense equivalent, and they share many properties. For example, the core of a game and its dual are equal. For more details on cooperative game duality, see for instance (Bilbao 2000).

Subgames


Let  S \subsetneq N be a non-empty coalition of players. The subgame  v_S : 2^S \to \mathbb{R} on  S is naturally defined as

 v_S(T) = v(T), \forall~ T \subseteq S.\,

In other words, we simply restrict our attention to coalitions contained in  S . Subgames are useful because they allow us to apply solution concepts defined for the grand coalition on smaller coalitions.

Properties for characterization

Superadditivity

Characteristic functions are often assumed to be superadditive (Owen 1995, p. 213). This means that the value of a union of disjoint coalitions is no less than the sum of the coalitions' separate values:

 v ( S \cup T ) \geq v (S) + v (T) whenever  S, T \subseteq N satisfy  S \cap T = \emptyset .

Monotonicity

Larger coalitions gain more:  S \subseteq T \Rightarrow v (S) \le v (T) . This follows from superadditivity if payoffs are normalized so singleton coalitions have value zero.

Properties for simple games

A coalitional game v is simple if payoffs are either 1 or 0, i.e., coalitions are either "winning" or "losing". Equivalently, a simple game can be defined as a collection W of coalitions, where the members of W are called winning coalitions, and the others losing coalitions. It is sometimes assumed that a simple game is nonempty or that it does not contain an empty set. In other areas of mathematics, simple games are also called hypergraphs or Boolean functions (logic functions).

A few relations among the above axioms have widely been recognized, such as the following (e.g., Peleg, 2002, Section 2.1[2]):

More generally, a complete investigation of the relation among the four conventional axioms (monotonicity, properness, strongness, and non-weakness), finiteness, and algorithmic computability[3] has been made (Kumabe and Mihara, 2011[4]), whose results are summarized in the Table "Existence of Simple Games" below.

Existence of Simple Games[5]
Type Finite Non-comp Finite Computable Infinite Non-comp Infinite Computable
1111 no yes yes yes
1110 no yes no no
1101 no yes yes yes
1100 no yes yes yes
1011 no yes yes yes
1010 no no no no
1001 no yes yes yes
1000 no no no no
0111 no yes yes yes
0110 no no no no
0101 no yes yes yes
0100 no yes yes yes
0011 no yes yes yes
0010 no no no no
0001 no yes yes yes
0000 no no no no

The restrictions that various axioms for simple games impose on their Nakamura number are also studied extensively.[6] In particular, a computable simple game without a veto player has a Nakamura number greater than 3 only if it is proper and non-strong.

Relation with non-cooperative theory

Let G be a strategic (non-cooperative) game. Then, assuming that coalitions have the ability to enforce coordinated behaviour, there are several cooperative games associated with G. These games are often referred to as representations of G. The two standard representations are:[7]

Solution concepts

The main assumption in cooperative game theory is that the grand coalition  N will form. The challenge is then to allocate the payoff  v(N) among the players in some fair way. (This assumption is not restrictive, because even if players split off and form smaller coalitions, we can apply solution concepts to the subgames defined by whatever coalitions actually form.) A solution concept is a vector  x \in \mathbb{R}^N that represents the allocation to each player. Researchers have proposed different solution concepts based on different notions of fairness. Some properties to look for in a solution concept include:

An efficient payoff vector is called a pre-imputation, and an individually rational pre-imputation is called an imputation. Most solution concepts are imputations.

The stable set

The stable set of a game (also known as the von Neumann-Morgenstern solution (von Neumann & Morgenstern 1944)) was the first solution proposed for games with more than 2 players. Let  v be a game and let  x ,  y be two imputations of  v . Then  x dominates  y if some coalition  S \neq \emptyset satisfies  x_i > y _i, \forall~ i \in S and  \sum_{ i \in S } x_i \leq v(S) . In other words, players in  S prefer the payoffs from  x to those from  y , and they can threaten to leave the grand coalition if  y is used because the payoff they obtain on their own is at least as large as the allocation they receive under  x .

A stable set is a set of imputations that satisfies two properties:

Von Neumann and Morgenstern saw the stable set as the collection of acceptable behaviours in a society: None is clearly preferred to any other, but for each unacceptable behaviour there is a preferred alternative. The definition is very general allowing the concept to be used in a wide variety of game formats.

Properties

The core

Main article: Core (economics)

Let  v be a game. The core of  v is the set of payoff vectors

 C( v ) = \left\{ x \in \mathbb{R}^N: \sum_{ i \in N } x_i = v(N); \quad \sum_{ i \in S } x_i \geq v(S), \forall~ S \subseteq N \right\}.\,

In words, the core is the set of imputations under which no coalition has a value greater than the sum of its members' payoffs. Therefore, no coalition has incentive to leave the grand coalition and receive a larger payoff.

Properties

The core of a simple game with respect to preferences

For simple games, there is another notion of the core, when each player is assumed to have preferences on a set X of alternatives. A profile is a list p=(\succ_i^p)_{i \in N} of individual preferences \succ_i^p on X. Here x \succ_i^p y means that individual i prefers alternative x to y at profile p. Given a simple game v and a profile p, a dominance relation \succ^p_v is defined on X by x \succ^p_v y if and only if there is a winning coalition S (i.e., v(S)=1) satisfying x \succ_i^p y for all i \in S. The core C(v,p) of the simple game v with respect to the profile p of preferences is the set of alternatives undominated by \succ^p_v (the set of maximal elements of X with respect to \succ^p_v):

x \in C(v,p) if and only if there is no y\in X such that y \succ^p_v x.

The Nakamura number of a simple game is the minimal number of winning coalitions with empty intersection. Nakamura's theorem states that the core C(v,p) is nonempty for all profiles p of acyclic (alternatively, transitive) preferences if and only if X is finite and the cardinal number (the number of elements) of X is less than the Nakamura number of v. A variant by Kumabe and Mihara states that the core C(v,p) is nonempty for all profiles p of preferences that have a maximal element if and only if the cardinal number of X is less than the Nakamura number of v. (See Nakamura number for details.)

The strong epsilon-core

Because the core may be empty, a generalization was introduced in (Shapley & Shubik 1966). The strong  \varepsilon -core for some number  \varepsilon \in \mathbb{R} is the set of payoff vectors

 C_\varepsilon( v ) = \left\{ x \in \mathbb{R}^N: \sum_{ i \in N } x_i = v(N); \quad \sum_{ i \in S } x_i \geq v(S) - \varepsilon, \forall~ S \subseteq N \right\}.

In economic terms, the strong  \varepsilon -core is the set of pre-imputations where no coalition can improve its payoff by leaving the grand coalition, if it must pay a penalty of  \varepsilon for leaving. Note that  \varepsilon may be negative, in which case it represents a bonus for leaving the grand coalition. Clearly, regardless of whether the core is empty, the strong  \varepsilon -core will be non-empty for a large enough value of  \varepsilon and empty for a small enough (possibly negative) value of  \varepsilon . Following this line of reasoning, the least-core, introduced in (Maschler, Peleg & Shapley 1979), is the intersection of all non-empty strong  \varepsilon -cores. It can also be viewed as the strong  \varepsilon -core for the smallest value of  \varepsilon that makes the set non-empty (Bilbao 2000).

The Shapley value

Main article: Shapley value

The Shapley value is the unique payoff vector that is efficient, symmetric, additive, and assigns zero payoffs to dummy players. It was introduced by Lloyd Shapley (Shapley 1953). The Shapley value of a superadditive game is individually rational, but this is not true in general. (Driessen 1988)

The kernel

Let  v : 2^N \to \mathbb{R} be a game, and let  x \in \mathbb{R}^N be an efficient payoff vector. The maximum surplus of player i over player j with respect to x is

 s_{ij}^v(x) = \max \left\{ v(S) - \sum_{ k \in S } x_k : S \subseteq N \setminus \{ j \}, i \in S \right\},

the maximal amount player i can gain without the cooperation of player j by withdrawing from the grand coalition N under payoff vector x, assuming that the other players in i's withdrawing coalition are satisfied with their payoffs under x. The maximum surplus is a way to measure one player's bargaining power over another. The kernel of v is the set of imputations x that satisfy

for every pair of players i and j. Intuitively, player i has more bargaining power than player j with respect to imputation x if s_{ij}^v(x) > s_{ji}^v(x), but player j is immune to player i's threats if  x_j = v(j) , because he can obtain this payoff on his own. The kernel contains all imputations where no player has this bargaining power over another. This solution concept was first introduced in (Davis & Maschler 1965).

The nucleolus

Let  v : 2^N \to \mathbb{R} be a game, and let  x \in \mathbb{R}^N be a payoff vector. The excess of  x for a coalition  S \subseteq N is the quantity  v(S) - \sum_{ i \in S } x_i ; that is, the gain that players in coalition  S can obtain if they withdraw from the grand coalition  N under payoff  x and instead take the payoff  v(S) .

Now let  \theta(x) \in \mathbb{R}^{ 2^N } be the vector of excesses of  x , arranged in non-increasing order. In other words,  \theta_i(x) \geq \theta_j(x), \forall~ i < j . Notice that  x is in the core of  v if and only if it is a pre-imputation and  \theta_1(x) \leq 0 . To define the nucleolus, we consider the lexicographic ordering of vectors in  \mathbb{R}^{ 2^N } : For two payoff vectors  x, y , we say  \theta(x) is lexicographically smaller than  \theta(y) if for some index  k , we have  \theta_i(x) = \theta_i(y), \forall~ i < k and  \theta_k(x) < \theta_k(y) . (The ordering is called lexicographic because it mimics alphabetical ordering used to arrange words in a dictionary.) The nucleolus of  v is the lexicographically minimal imputation, based on this ordering. This solution concept was first introduced in (Schmeidler 1969).

Although the definition of the nucleolus seems abstract, (Maschler, Peleg & Shapley 1979) gave a more intuitive description: Starting with the least-core, record the coalitions for which the right-hand side of the inequality in the definition of  C_\varepsilon( v ) cannot be further reduced without making the set empty. Continue decreasing the right-hand side for the remaining coalitions, until it cannot be reduced without making the set empty. Record the new set of coalitions for which the inequalities hold at equality; continue decreasing the right-hand side of remaining coalitions and repeat this process as many times as necessary until all coalitions have been recorded. The resulting payoff vector is the nucleolus.

Properties

Convex cooperative games

Introduced by Shapley in (Shapley 1971), convex cooperative games capture the intuitive property some games have of "snowballing". Specifically, a game is convex if its characteristic function  v is supermodular:

 v( S \cup T ) + v( S \cap T ) \geq v(S) + v(T), \forall~ S, T \subseteq N.\,

It can be shown (see, e.g., Section V.1 of (Driessen 1988)) that the supermodularity of  v is equivalent to

 v( S \cup \{ i \} ) - v(S) \leq v( T \cup \{ i \} ) - v(T), \forall~ S \subseteq T \subseteq N \setminus \{ i \}, \forall~ i \in N;\,

that is, "the incentives for joining a coalition increase as the coalition grows" (Shapley 1971), leading to the aforementioned snowball effect. For cost games, the inequalities are reversed, so that we say the cost game is convex if the characteristic function is submodular.

Properties

Convex cooperative games have many nice properties:

Similarities and differences with combinatorial optimization

Submodular and supermodular set functions are also studied in combinatorial optimization. Many of the results in (Shapley 1971) have analogues in (Edmonds 1970), where submodular functions were first presented as generalizations of matroids. In this context, the core of a convex cost game is called the base polyhedron, because its elements generalize base properties of matroids.

However, the optimization community generally considers submodular functions to be the discrete analogues of convex functions (Lovász 1983), because the minimization of both types of functions is computationally tractable. Unfortunately, this conflicts directly with Shapley's original definition of supermodular functions as "convex".

See also

References

  1. 2^N denotes the power set of N.
  2. Peleg, B. (2002). "Chapter 8 Game-theoretic analysis of voting in committees". Handbook of Social Choice and Welfare Volume 1. Handbook of Social Choice and Welfare 1. pp. 195–201. doi:10.1016/S1574-0110(02)80012-1. ISBN 9780444829146.
  3. See a section for Rice's theorem for the definition of a computable simple game. In particular, all finite games are computable.
  4. Kumabe, M.; Mihara, H. R. (2011). "Computability of simple games: A complete investigation of the sixty-four possibilities" (PDF). Journal of Mathematical Economics 47 (2): 150–158. doi:10.1016/j.jmateco.2010.12.003.
  5. Modified from Table 1 in Kumabe and Mihara (2011). The sixteen Types are defined by the four conventional axioms (monotonicity, properness, strongness, and non-weakness). For example, type 1110 indicates monotonic (1), proper (1), strong (1), weak (0, because not nonweak) games. Among type 1110 games, there exist no finite non-computable ones, there exist finite computable ones, there exist no infinite non-computable ones, and there exist no infinite computable ones. Observe that except for type 1110, the last three columns are identical.
  6. Kumabe, M.; Mihara, H. R. (2008). "The Nakamura numbers for computable simple games". Social Choice and Welfare 31 (4): 621. doi:10.1007/s00355-008-0300-5.
  7. Aumann, Robert J. "The core of a cooperative game without side payments." Transactions of the American Mathematical Society (1961): 539-552.

Further reading

External links

This article is issued from Wikipedia - version of the Wednesday, April 20, 2016. The text is available under the Creative Commons Attribution/Share Alike but additional terms may apply for the media files.