Information Technology Reference
In-Depth Information
Later, two related greedy algorithm-based heuristics (one determinis-
tic, one not) were proposed in [14]. However, stability was not handled,
and social utility was used instead. Then in [15] a round-based mecha-
nism was proposed for TU games, where the agents, arranged in some
random order, are allowed to make proposals to other agents, and
coalitions are formed when they are accepted by all agents. The au-
thors used social utility as the evaluation criterion in their experiments.
Kernel was later used in an enhanced version of their mechanism [16].
Other coalition formation mechanisms have been proposed for spe-
cific applications. For example, in a buyer coalition mechanism, an ap-
proach targeting TU-core-based solutions was proposed in [13]. Later,
in 2006, a centralized mechanism was proposed in [19], where the au-
thors used Pareto optimality as measurement of their solution qual-
ity. More recently, a distributed randomized mechanism approach has
been proposed, in which the evaluation criteria used was both Pareto
optimality and the NTU core [20].
A summary is given in Table 2.1.
Tab l e 2 . 1 Selected coalition formation mechanisms and their solution
concepts
Game
Type
Common
Knowledge
Mechanism
Approach
Solution Concept
[10]
Auction
TU
Core
Yes
[11]
Propose-and-Evaluate
TU
Kernel
Yes
[12]
Propose-and-Evaluate
TU
Core
Yes
[13]
Greedy Algorithm
TU
Social utility and Core
Yes
[14]
Greedy Algorithm
TU
Social utility
Yes
[15]
Propose-and-Evaluate
TU
Social utility
No
[16]
Propose-and-Evaluate
TU
Social utility and Kernel
No
[17]
Propose-and-Evaluate
TU
Core
Yes
[18]
Auction
TU
Social utility
Yes
[19]
Centralised Search
NTU
Pareto Optimality
Yes
[20]
Propose-and-Evaluate NTU Pareto Optimality and Core
Yes
Search WWH ::




Custom Search