Many multiagent systems are comprised of self-interested, autonomous
agents that have to collaborate with each other in order to maximize
their individual utilities. Most real world agents are inhibited with
limited resources, capabilities and knowledge about their environment,
and their interactions are limited to a select set of other agents
depending on their social network. Typically, such systems also lack
any central authority to resolve issues between multiple agents. Such
agents can pool their resources and combine their services to handle
complex tasks, which they would otherwise be unable to accomplish, by
forming coalitions. Coalition formation entails determining solutions
that are stable. Some of the other desirable properties for multiagent
systems are social welfare maximization and fairness of revenue
distribution amongst the agents. Solving some of these properties are
computationally expensive as they involve a combinatorial search, and
considering them together exacerbates the problem further. Our
research focus is on devising bargaining processes that allow the
agents to act in their best interests, while the system as a whole
reaches an equilibrium state that is utilitarian, fair, and stable. In
this paper, we present our negotiation algorithm where agents make
their decisions based only on their local information, and handle
tasks in a completely distributed fashion. We present experimental
results to demonstrate the quality of our solution in terms of agent
utility maximization, stability in terms of the bargaining set, and
scalability for various network topologies.