Example Application of the b-Core: Buyer Coalition Problem with Beliefs - Belief-based Stability in Coalition Formation with Uncertainty - page 118

5.1.2 How About Stability?

The second concern is stability . In buyer coalition problems, as in other
semi-competitive games, the stability of coalitions is more important
than social utility. In short, a stable solution is one that no agents
would have the incentive to deviate from the agreed buyer coalitions.
Again, there are two reasons that coalition stability is more suitable
as solution concepts for group buying problem.
First, buyer agents, like any other types of agents, are assumed to
be rational so that they are selfish and interested in their own utility
only. That is, their only concern is that whether there exist some alter-
nate solutions where they (either alone or as a sub-group) are better
off, regardless the social utility. If there is one such alternate solution,
then they will definitely leave for the better coalition, otherwise, they
are happy to stay.
Second, we cannot force any agents to sacrifice for the sake of
global optimality, a point that is either not considered in most existing
approaches, or only given limited handling. An unstable coalition is
still meaningless in real life even if it can achieve a good social utility.

5.1.3 Private Information in Buyer Coalition

The third concern is privacy of information . Buyer agents are privately
owned software tools, where personal information is stored as private
beliefs of the agents. Therefore, a buyer agent's preference order for
some items should be private and known to that agent only, and he
may not be willing to disclose these preferences.
However, most of the approaches so far have a common knowledge
assumption in that each individual agent's preference are either known
to all agents (or at least known by a central agent that computes the
solution), or that the agents are willing to disclose such information