The Core

Allocations no coalition can profitably abandon.

The idea

The core is the solution concept for coalitional games: the set of divisions of the total payoff that no group of players can improve on by breaking away.

A coalitional game with transferable payoff is a pair $(N, v)$: a set $N$ of $n$ players, and a function $v$ assigning each coalition $S \subseteq N$ a worth $v(S)$, the total payoff the members of $S$ can guarantee themselves with no help from anyone outside. Suppose all the players cooperate and produce $v(N)$. An allocation $x = (x_{1}, \ldots, x_{n})$ is feasible when it distributes exactly what was produced, $\sum_{i} x_{i} = v(N)$. Feasibility does not settle the division: a coalition receiving less than its own worth can secede and do better.

The core is the set of feasible allocations satisfying

$\sum_{i \in S} x_{i} \ge v(S) \quad \text{for every coalition } S \subseteq N.$

Every coalition then receives at least what it could produce by itself, so seceding gains it nothing. The deviating unit is a group, not a single player.

The core can be empty: when the coalitions' worths add up to more than $v(N)$, no feasible allocation meets every constraint. When it is nonempty, it is usually a region rather than a single point, because the definition selects the stable divisions, not a unique fair one. The figure shows the three-player picture: the feasible allocations form a triangle, each pairwise constraint $\sum_{i \in S} x_{i} \ge v(S)$ cuts a strip off one corner, and the core is the region that survives every cut.

Ways to work on it

Builds on

Leads to

Not sure where to start? Take the ten-question placement test.