Artificial Neural Network-Based Research On TDMA Broadcast Scheduling In WMNs

Wireless multihop networks (WMNs) has been widely used in many fields for it can beused to provide wirless data communication services among a numbers of nodes dispersedover a broad geographic region. In WMNs, untrolled data transmissions among the nodesoften cause conflicts. In order to avoid conflicts, the time division multiple access (TDMA)has been adopted to solve the TDMA broadcast scheduling problem (BSP) in WMNs. TheTDMA BSP of WMNs has been proven to be NP-complete, and artificial neuralnetwork-based methods have been proven to be efficient for BSP.The optimization performance of Sun-noise-tuning-based hysteretic noisy chaotic neuralnetwork (Sun-NHNCNN) is more excellent than that of all other neural networks in solvingTDMA BSP of WMNs. However, the internal noises of Sun-NHNCNN cannot be controlledcompletely, and the optimization performance of Sun-NHNCNN would be weakened if thenoise amplitudes were higher. Therefore, as the initial noise amplitudes are higher, theoptimization performance of Sun-NHNCNN would not be as good as that as the initial noiseamplitudes are lower, and for this reason, the application of Sun-NHNCNN would be limited.In order to improve the optimization performance of Sun-NHNCNN as the initial noiseamplitudes are higher, in this paper, a new noise-tuning factor is introduced into the outputitem of Sun-NHNCNN, and an improved hysteretic noisy chaotic neural network (IHNCNN)is proposed. The internal noises of newly proposed IHNCNN can be controlled completely bytuning the values of the newly introduced noise-tuning factor and the original noise-tuingfactor of Sun-NHNCNN. Therefore, the optimization performance of the newly proposedIHNCNN would be more excellent than NHNCNN as the initial noise amplitudes are higher.In the process of maximizing the time slot utilization of TDMA frame, the adoptedenergy function of current continuous neural networks often leads the final TDMA frame tobe infeasible. In order to overcome the above-mentioned weakness, in this paper, a constraintitem of preventing data transmission conflicts in WMNs is introduced into the optimizationprocess, and by this way, the validity of final TDMA frame can be guaranteed.The IHNCNN and the modified method of maximizing the time slot utilization ofTDMA frame are applied to sovle the TDMA BSP of WMNs. The simulation results showthat the proposed methods in this paper are effectiveness. In the last of this paper, the TDMA BSP of WMNs with moving nodes would beanalyzed and researched, and a valid algorithm with specific solving steps for the problem isproposed. The effectiveness of the proposed algorithm can be demonstrated by its succeededapplying on solving the TDMA BSP of WMNs with moving nodes.