Multi-agent based execution environment for task allocation via coalition formation

Abstract

This paper focuses on a solution to the problem of task allocation through coalition formation and presents the design and implementation of a framework for this purpose. The framework serves as a multi-agent execution environment where worker agents follow the Shehory-Kraus’s coalition formation algorithm to negotiate and form coalitions. Worker agents prefer to form an optimal coalition to maximize the joint utility of the system as a whole, rather than to maximize their own utilities. There is no central authority for distribution of tasks among the agents. The framework provides the infrastructure, which, allows the formation of coalitions, assigns a task to each coalition and then monitors task execution, taking the necessary steps for rescheduling of tasks in case of noncompletion due to agent errors, generating and recording of task execution reports, handling cyclic execution of coalition formation process. In addition, an efficient and fast messaging infrastructure has been developed for effective agent communication.