Dynamic Channel Assignment for Large-Scale Cellular Networks Using Noisy Chaotic Neural Network

This paper presents a novel dynamic channel assignment (DCA) technique for large-scale cellular networks (LCNs) using noisy chaotic neural network. In this technique, an LCN is first decomposed into many subnets, which are designated as decomposed cellular subnets (DCSs). The DCA process is independently performed in every subnet to alleviate the signaling overheads and to apportion the DCA computational load among the subnets. Then a novel energy function is formulated to avoid causing mutual interference among neighboring subnets based on the real-time interference channel table. In each subnet, the proposed energy function also satisfies three interference constraints among cells and the number of required channels of each cell, and simultaneously minimizes the total number of assigned channels to improve spectrum utilization. A typical 441-cell LCN with 70 available channels, which can be decomposed into nine 49-cell DCSs, is examined to demonstrate the validity of the proposed technique by blocking probability, including uniform and hot spot traffic patterns.