Multi-agent system for radio resource management of a heterogeneous network

This paper presents a multi-agent hierarchy that takes advantages of a greedy algorithm in conjunction with a Taboo search method to solve the problem of distributing radio resources in a wireless network. By the maximum reuse of the same frequencies, the method allows sharing all resources between different RANs (Regional Radio Access Network), coordinating these RANs to facilitate grouping the frequencies in the lower part of the spectrum in order to allow a later new assignment and best improvement of neighborhood. The approach is based on a macro-level management taking the form of a hierarchical group of distributed agents in the network and handling all RANs in a localized region regardless of the operating band. Where each regional network is overseen by a supervisor agent and it associates an agent for each cell. Our approach has been tested on GRAPH benchmarks of frequency assignment of the CALMA project. The results obtained are equivalent to those of current methods but our approach shows more efficiency in terms of flexibility and autonomy.