New mixed broadcast scheduling approach using neural networks and graph coloring in wireless sensor network

Due to the mutual interference and sharing of wireless links in TDMA wireless sensor networks, conflicts will occur when data messages are transmitting between nodes. The broadcast scheduling problem (BSP) is aimed to schedule each node in different slot of fixed length frame at least once, and the objective of BSP is to seek for the optimal feasible solution, which has the shortest length of frame slots, as well as the maximum node transmission. A two-stage mixed algorithm based on a fuzzy Hopfield neural network is proposed to solve this BSP in wireless sensor network. In the first stage, a modified sequential vertex coloring algorithm is adopted to obtain a minimal TDMA frame length. In the second stage, the fuzzy Hopfield network is utilized to maximize the channel utilization ratio. Experimental results, obtained from the running on three benchmark graphs, show that the algorithm can achieve better performance with shorter frame length and higher channel utilizing ratio than other exiting BSP solutions.