Multi-agent Cooperation and Stability - Belief-based Stability in Coalition Formation with Uncertainty - page 7

Later a more general version of the game is proposed where the utilities
are no longer required to be transferable.

1.2.1 Transferable Utility Games

The idea of transferable utility cooperative games (TU games) was first
proposed in a topic by von Neumann and Morgenstern (1944), entitled
Theory of Games and Economic Behaviour [13]. Typically, there are
three steps in coalition formation in transferable utility games.
1) The formation of coalition structure
This is the step where the individual agents come together to form
exhaustive and non-overlapping coalitions.
2) Action coordination
The agents in the same coalition then agree to decide how to coordi-
nate their action so that their achieved utility as a group is maximized.
In game theory, this step is frequently assumed instead of explicitly
described.
3) Utility sharing
The agents need to reach agreement on how to divide the utility
amongst the members in such a way that every member considers it to
be 'fair' and thus have no desire to leave the coalition. Several trans-
ferable utility game solution concepts exist, with the most popular one
being the core [13]. However, there are also other solution concepts,
including the stable set [13], the nucleolus [14], the kernel [15], and the
Shapley value [12].
Several conditions must be met in order to apply a transferable
utility solution concept:
Firstly, as in social utility, the transferable utility-based solution
concepts require that the individual agent's utilities can be meaning-
fully added up. For example, if the utilities are defined as the profit
of the individual agents, profits must be able to be losslessly added