改进型WSN覆盖模型及其求解的果蝇视觉进化神经网络  

Improved WSN Coverage Model and Related Fly Visual Evolutionary Neural Network

作　　者：黄唯 张著洪 HUANG Wei;ZHANG Zhuhong(College of Big Data and Information Engineering,Guizhou University,Guiyang 550025,China;Guizhou Provincial Characteristic Key Laboratory of System Optimization and Scientific Computation,Guiyang 550025,China)

机构地区：[1]贵州大学大数据与信息工程学院,贵阳550025 [2]贵州省系统优化与科学计算特色重点实验室,贵阳550025

出　　处：《小型微型计算机系统》2025年第1期33-43,共11页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(62063002)资助。

摘　　要：传感器节点的随机部署易于导致WSN的网络覆盖率低和连通性差,进而影响WSN的服务质量;如何构建节点部署规划模型及探究其求解算法,仍然是WSN研究面临的科技难题.为此,提出改进型WSN覆盖优化模型及其求解的果蝇视觉进化神经网络优化算法.模型设计中,在已有覆盖率指标下,引入连通度指标以保证网络的连通性,进而借助正三角形法构建确保区域内节点均匀部署的约束限制条件,获得以覆盖率和连通度的加权和为性能指标的改进型WSN覆盖优化模型.算法设计中,依据注意力和果蝇视觉系统的信息处理机制,获得能处理约束条件且能输出全局和局部学习率的改进型果蝇视觉神经网络,进而将其输出与基于改进型蜣螂优化的状态更新策略组合,获得能处理强非线性约束优化及WSN覆盖优化问题的改进型果蝇视觉进化神经网络优化算法.比较性的实验结果显示,所获算法不仅具有强的竞争力,而且也暗示视觉信息处理机制与元启发式方法结合对解决约束优化问题具有较好潜力.The random deployment of sensor nodes in WSN easily causes the problem of the low network coverage and network connectivity,which directly influences the WSN′s service quality.It is still a scientific and technological challenge to construct sensor node deployment planning models and probe into the related approaches.Therefore,this work develops an improved WSN coverage optimization model and the related fly visual evolutionary neural network optimization approach.In the design of the model,the connectivity index is used to ensure the connectivity of the network,while the regular triangle method is employed to construct a constraint condition so that the uniform deployment of nodes in the region can be achieved.Afterwards,a single-objective constrained optimization model is derived based on the weighed coefficient approach.In the design of the algorithm,an improved fly visual evolutionary neural network optimization algorithm is proposed to handle constrained optimization problems and WSN coverage problems.Therein,an improved fly visual neural network able to handle constraints is developed to output global and local learning rates,based on the two mechanisms of visual attention and visual information-processing of the fly visual system;a state update strategy is constructed to update current states,after the basic dung beetle optimization approach is improved by means of the Levy flight strategy and a dynamically adjustable nonlinear boundary selection factor.The comparative experiments have validated that,not only the acquired optimization approach is strongly competitive,but also the integration of the visual information processing mechanism and metaheuristics is of great potential to solving constrained optimization problems.

关 键 词：无线传感器网络覆盖优化 网络连通度 果蝇视觉神经网络 视觉进化神经网络 蜣螂优化 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]