Cooperative negotiation in a multi-agent system for real-time load balancing of a mobile cellular network

Abstract

A cooperative negotiation approach for the real-time control of cellular network coverage is described. The performance of the whole cellular network is improved by contracting and shaping the antenna radiation pattern around traffic hot spots and expanding adjacent cell coverage to fill in the coverage loss. The paper shows that the local area real time cooperative negotiation between base stations leads to a near global optimal coverage agreement which is reached in the context of the whole cellular network. Results showing the advantage of this technique are presented. Global optimization using constrained real-coded genetic algorithms provides the benchmark. In our work, instead of using a formal negotiation model of alternating offers, we create certain number of possible local hypotheses and start negotiations based on them. Some negotiation may reach agreements before their deadlines, and the system commits to the best agreement found at the end. This approach is more predictable and controllable than the formal negotiation model. Architecturally the negotiation component described is part of the planning layer of an agent system for resource management in 3G wireless networks. This has a layered architecture with both planning and reactive components. The results are allowing the development of a novel geographic load-balancing scheme for cellular networks that intelligently changes cellular coverage according to the geographic traffic distribution in real time.