Information Technology Reference
In-Depth Information
possible procedures for leader election [16], though in order of placing into focus our
mechanism we have chosen a straightforward approach of assigning this role for each
coalition to the PA with the highest association coefficient.
We note in advance that the process executes asynchronously and in parallel for
all agents. Communication amongst agents assumes the use of time-outs by means of
which agents place upper bounds, specifying the amount of time allocated for receiving
a reply. In case no reply is received in due time, the particular agent is simply disre-
garded from being considered as a candidate for coalition reorganization. Given the
classification of Section 2, our proposed approach belongs to the class of decentralized
and dynamic methods. Here coalition adaptation is achieved in a self-organizing fash-
ion by opportunistic aggregation of agents, while maximizing coalitional benefits by
means of taking advantage of local network resources.
As it had been described in Section 3, all actors
a
i
submit on a daily basis their fore-
casted profile
β
i
, which typically does not differ exceedingly from their previous one.
Nevertheless these cumulative variations might entail a reorganization of the coalition
structure
CS
for the following forcasted period in order to assure enhanced coordination
at the coalition level. Therefore, consequent to calculating the energetic balance
φ
2
of
the coalition given the existing
LA
sand
PA
s comprising it, it is to be determined the
PA
actors that would qualify to be signed-off, or the profile of the actors that would
be eligible to be signed-in to the coalition. The association coefficients
γ
i,j
, revealing
the existing interdependencies within the coalition, play a key role at this stage. The
weakest links signify the actors the coalition is least dependent on, based on which
agents are to be proposed for being signed out of a coalition. Interconnected coalitions
should incorporate a control mechanism for achieving a basic energy optimization for
the entire system via a close coordination with neighboring coalitions. Otherwise, the
security and stability of the main grid could be threatend severely. The mechanism is in
such a way designed that it proves to be consistent with our proposed solution concept
introduced hereafter.
Thus, the problem we are facing in open organizations requires a modification of the
coalition structure due to the variations occurring within the system. With these consid-
erations in mind we seek a notion of equilibrium that intrinsically provides an argumen-
tation scheme, which allows for a reorganization of the coalition structure. Furthermore,
the solution concept should reflect the decentralization outlook of our scenario, while
minimizing the structural adaptations by providing a minimum set of interaction rules
in order of attaining the desired stability properties amongst negotiating agents.
The solution we propose can be directly referenced to game-theoretic approaches
on issues of stability and negotiation. For further considerations on notions of sta-
bility, their strength, limitations and interrelations we refer the reader to [9]. In our
scenario, of utmost importance is the agents' capability to coordinate and reorganize
into groups or coalitions by transforming traditional game-theory criterions of stabil-
ity towards operating in dynamic environments. Moreover, we advocate for reorienting
game-theory to accomodate situations where coordination is a more likely descriptor of
the game rather than simply self-interested settings. As it is emphasized in [9], an equiv-
alent formulation for solution concepts can be in interpreted in terms of objections and
