Investigating adaptive, confidence-based strategic negotiations in complex multiagent environments

摘要

We propose an adaptive 1-to-many negotiation strategy for multiagent coalition formation in complex environments that are dynamic, uncertain, and real-time. Our strategy deals with how to assign multiple issues to a set of concurrent negotiations based on an initiating agent’s confidence in its profiling of its peer agents. When an agent is confident, it uses a packaged approach—conducting multiple multi-issue negotiations—with its peers. Otherwise, it uses a pipelined approach—conducting multiple single-issue negotiations—with its peers. The initiating agent is also capable of using both approaches in a hybrid, dealing with a mixed group of responding peers. An agent’s confidence in its profile or view of another agent is crucial, and that depends on the environment in which the agents operate. To evaluate the proposed strategy, we use a coalition formation framework in a complex environment. Results show that the proposed strategy outperforms the purely pipelined strategy and the purely packaged strategy in both efficiency and effectiveness.