Quality of service and grade of service optimisation with distributed DCA schemes based on Hopfield neural network algorithm
D. Girma, Óscar Lázaro · 2002
Dynamic channel allocation (DCA) schemes based on the Hopfield neural network (HNN) have previously been reported to perform favourably with other competing schemes and provide optimum channel assignment in a short period of time. This paper presents a comparative study of centralised and distributed HNN-DCA schemes in terms of call blocking probability (on call admission) and call dropping probability (after call connection). The paper also addresses how a distributed optimisation area size (DOA), an important feature of the proposed distributed control, can be selected for achieving an optimum balance between call blocking and call dropping for improving an overall call capacity.