In the information collection work of the wireless sensor network,it is required to add location information of the sensor node to ensure the reliability of the information.Therefore,how to obtain the location information of unknown nodes in the network and reduce the positioning error of the node location algorithm are the main considerations for the node location problem of wireless sensor networks.The Flower Pollination Algorithm(FPA)is a new type of heuristic group intelligent optimization algorithm.Compared with other algorithms,FPA algorithm has better global search and local search balance,but it also has the problems of lack of mutation mechanism and easy to fall into local optimum.Aiming at these problems,an improved flower pollination algorithm(IFPA)is proposed.The IFPA algorithm adds a weighting factor to the Levy flight step in the global pollination process.The weighting factor changes according to the number of iterations of the algorithm,which speeds up the convergence of the algorithm.In the local pollination process,the historical optimal solution is introduced to make the pollen individual positionally.Update,while adding Gaussian mutation factor as a mutation mechanism,strengthens the ability of the algorithm to jump out of local optimum.Using four classical standard test functions,the IFPA algorithm is compared with the Particle Swarm Optimization(PSO)and FPA algorithms to verify that the IFPA algorithm converges faster and has higher precision.This paper studies the application of IFPA algorithm to wireless sensor network location,which mainly includes the following two aspects: First,for the traditional RSSI positioning algorithm,the least squares method has a great influence on the accuracy of the unknown node location.The IFPA-RSSI positioning algorithm is proposed.In the presence of ranging error,the RSSI positioning problem is transformed into the problem of minimizing the ranging error.The IFPA algorithm is used to solve the unknown node position and improve the positioning accuracy of the RSSI positioning algorithm.The basic idea of the algorithm is to set the initial pollen individual group,each pollen individual represents a candidate solution of the unknown node position,and the error between the actual distance from the unknown node to the anchor node and the distance measurement is used as the fitness function to perform the pollen individual.The position is iteratively updated to arrive at the optimal value,which is the coordinate value closest to the actual position of theunknown node.Experimental simulations show that the positioning accuracy of IFPA-RSSI algorithm is better than traditional RSSI positioning algorithm,PSO-RSSI positioning algorithm and FPA-RSSI positioning algorithm.Secondly,for the problem that the positioning accuracy of the ELM positioning algorithm is greatly affected by the input weight and the implicit threshold initial assignment in the ELM network.The IFPA algorithm is used to improve the ELM positioning algorithm,and the IFPA-ELM positioning algorithm is proposed.The basic idea of the algorithm is to use the IFPA algorithm as the learning method of the ELM network.The error between the network output value and the expected output value is used as the fitness function to optimize the initial input weight and the implicit threshold of the ELM network.The initial structure of the ELM network is assigned according to the optimal value,and the optimized ELM network is used to solve the unknown node position.Experimental simulations show that the positioning accuracy of IFPA-ELM positioning algorithm is better than traditional ELM positioning algorithm,PSO-ELM based positioning algorithm and FPA-ELM based positioning algorithm. |