Reinforcement-learning-based self-organisation for cell configuration in multimedia mobile networks

In future wireless code division multiple access (WCDMA) cellular networks, random user mobility and time-varying multimedia traffic activity make the system design of coverage and capacity become a challenging issue. To utilise radio resource efficiently, it is crucial for cellular networks to have the capability of self organisation for cell configuration, which can configure service coverage and system capacity dynamically to balance traffic loads among cells by being aware of the system situation. This paper proposes a reinforcement-learning-based self-organisation scheme for cell configuration in multimedia mobile networks, which takes into account both pilot power allocation and call admission control mechanisms. Simulation results show that the proposed scheme improves system performance significantly compared to the conventional fixed pilot power allocation scheme and the scheme in which only pilot power is adjusted dynamically but the criterion of the call admission control is not coupled to it. Copyright (c) 2005 AEIT.