Distributed Meeting Schedule through Cooperative Mediation:
analysing
OptAPO's performance in a complex scenario
Paulo R. Ferreira Jr. and Ana L. C. Bazzan
Abstract
In multiagent systems, Distributed Constraint
Optimization
Problem (DCOP) has been used as a formalism to model a wide range of
agents
coordination issues. An approach based on cooperative me-diation was
recently
proposed as a new way to find the optimal solu-tion to DCOP. An
important
question about any solution to DCOP is whether it is fast enough to be
applied
in real-world applications, such as Distributed Meeting Scheduling
problems.
Here, we map the DMS as a DCOP, use cooperative mediation, and compare
the
performance of this approach with the results achieved by another
algorithms
for DCOP. Given the time complexity of the complete solutions, we
propose
a modified approach for the cooperative mediation, in which the idea is
to
trade the completeness of the search mechanism for the performance of a
heuristic
search, which yields a good solution in a feasible time.