The boundary of the core of a balanced game: face games (Q776857): Difference between revisions

A coalition \(S\) of a transferable utility game \((N,v)\) is exact if some core element \(x\) is effective for this coalition, i.e., \(\sum_{i\in S} x_i=v(S)\). The game is exact if all coalitions are exact. The game that may differ from \((N,v)\) at most inasmuch as it assigns the amount that \(S\) cannot be prevented from by its complement, i.e., \(v(N)-v(N\setminus S)\), to \(S\), is called ``\(S\)-cut'' game. The core of this \(S\)-cut game is nonempty if and only if \(N\setminus S\) is exact. In this case, the \emph{\(S\)-face} game is the unique exact game the core of which coincides with the core of the \(S\)-cut game. It is shown that the definition of the face game expands the definition introduced by \textit{J. González-Díaz} and \textit{E. Sánchez-Rodríguez} [Games Econ. Behav. 62, No. 1, 100--105 (2008; Zbl 1135.91315)] for a convex game. The authors prove that the core of the game is the convex hull of the cores of its face games. Moreover, several relations between the core of the game and the cores of the face games are deduced. For instance, it turns our that (a) a marginal contribution vector (assigning, consecutively according to some ordering, the marginal contribution to each player) is a core element if it belongs to the core of some singleton-face game and (b) a lexicographically maximal core element is a lexicographically maximal element of the core of a some singleton-face game.