Research On Hybrid Location Algorithm For Wireless Sensor Networks Based On Differential Evolution

Wireless Sensor Network(WSN)is a technology with low cost,low energy consumption and strong adaptability,which has been widely used in real life.Node localization technology plays a crucial role in wireless sensor networks,helping the entire system achieve precise localization.Due to the unpredictable distribution of sensor nodes,effective collaborative cooperation must be adopted to ensure that they can be accurately located to their destination.How to develop more effective and accurate localization algorithms and more accurate node position estimation methods,It has become a hot topic in the field of WSN research today.In order to improve positioning accuracy without increasing hardware costs,this paper adopts a strategy of hybrid optimization of differential evolution and other intelligent algorithms,which is introduced into the field of wireless sensor node localization.The main tasks are as follows:In order to solve the problems of premature convergence and slow optimization of differential evolution algorithm,a differential adaptive evolution particle swarm optimization(SA-DEPSO)based on particle swarm optimization is proposed.Firstly,the localization problem is transformed into a function iterative optimization problem using the least squares method.Secondly,based on particle swarm optimization algorithm,hybrid adaptive differential evolution strategy is used.The hybrid algorithm can not only avoid the problem of premature convergence,but also improve the optimization speed and reduce positioning errors.The experimental simulation results show that compared with the DE-PSO,SA-MCDE,and PSO-ALS algorithms,the proposed SA-DEPSO algorithm reduces the average number of runs by 55,45,and25,respectively;The average positioning error decreased by 17.3%,13.1%,and 7.5%,respectively.The proposed algorithm is faster in optimization and has smaller errors.In order to solve the problem of lacking mutation mechanism and easily falling into local optima in flower pollination algorithms,a Cauchy Elite Flower Pollination Algorithm(CE-FPA)is proposed,which effectively integrates Cauchy distribution and elite opposition learning.Firstly,the elite opposition strategy is introduced to enhance population diversity while directing the approximate solution of the current solution space to the global optimal solution space,thereby improving the search ability of the algorithm;Then,the Levy flight step used in the original pollination process is replaced with a Cauchy perturbation factor.The Cauchy mutation operator can improve the convergence ability of the flower pollination algorithm,and improve the stability of the algorithm while jumping out of local optima;Finally,by integrating these two strategies,the overall performance of the algorithm is improved.The CE-FPA algorithm was compared with FPA-PSO,AFPA,pc-FPA,and FPA algorithms through eight standard test functions in multidimensional experiments.The simulation results showed that the CE-FPA algorithm has the advantages of faster convergence and higher accuracy.