Multi-agent Cooperation and Stability - Belief-based Stability in Coalition Formation with Uncertainty - page 8

up (e.g., the money should be freely transferable from one agent to
another without loss, where it is added up.)
Secondly, as mentioned above, they must be a commonly agreed
upon mechanism to fairly redistribute the utilities, again losslessly,
amongst the agents. The idea is that even if some of the individual
agents may suffer a loss or less than desired profit, the total profit
can still be redistributed in such a way that no agent will have the
incentive to leave.
Thus, the transferable utility solution concepts assume that the
utility gained by any agent is freely transferable to other agents, but
this is not always the case. For example, imagine a problem where the
utilities represent the agents' satisfaction instead of some transferable
monetary values. As most people's degree of satisfaction is subjected,
and clearly cannot be transferred. The usefulness of a transferable
utility solution concepts for such problems is questionable.

1.2.2 Non transferable Utility Games

As a result, in 1960, Aumann and Peleg investigated a more gen-
eral version of cooperative game without transferable utility [16]. The
resulting concepts are known as non-transferable utility cooperative
games (NTU games).
The idea of an NTU game is similar to that of a transferable utility
game except that the utilities are awarded to each individual instead of
to the coalitions as a whole (i.e, they do not need to be added up), and
they are no longer transferable among the members. As a result, while
agents in transferable utility games can settle any conflict of interests
via side-payments, this is no longer possible in non-transferable util-
ity games. The non-transferable utility core [17] and non-transferable
utility Shapley Value [18] are two of the more well known NTU games
solution concepts.
Game theoretical concepts such as the core can make sure that