An improved multi-agent simulation methodology for modelling and evaluating wireless communication systems resource allocation algorithms

摘要

Multi-Agent Systems (MAS) constitute a well known approach in modelling dynamical real world systems. Recently, this technology has been applied to Wireless Communication Systems (WCS), where efficient resource allocation is a primary goal, for modelling the physical entities involved, like Base Stations (BS), service providers and network operators. This paper presents a novel approach in applying MAS methodology to WCS resource allocation by modelling more abstract entities involved in WCS operation, and especially the concurrent network procedures (services). Due to the concurrent nature of a WCS, MAS technology presents a suitable modelling solution. Services such as new call admission, handoff, user movement and call termination are independent to one another and may occur at the same time for many different users in the network. Thus, the required network procedures for supporting the above services act autonomously, interact with the network environment (gather information such as interference conditions), take decisions (e.g. call establishment), etc, and can be modelled as agents. Based on this novel simulation approach, the agent cooperation in terms of negotiation and agreement becomes a critical issue. To this end, two negotiation strategies are presented and evaluated in this research effort and among them the distributed negotiation and communication scheme between network agents is presented to be highly efficient in terms of network performance. The multi-agent concept adapted to the concurrent nature of large scale WCS is, also, discussed in this paper.