Research On Improvements And Applications Of Coyote Optimization Algorithm

Coyote optimization algorithm(COA)is an intelligent optimization algorithm inspired by the behavior of coyote populations,which simulates the growth,birth,death,and migration of coyotes.In this algorithm model,the population is evenly divided into fixed groups of coyotes in a certain proportion,and the growth of individual coyotes is influenced by the cultural trends of other wolves in the group to which they belong,as well as the original growth state of the coyote.The birth of coyote individuals is to randomly select two coyotes to produce pups according to the mating rules.The death of coyotes follows the natural law of birth,aging,disease and death to eliminate coyotes.The migration of coyotes is that,under a certain probability,coyotes leave the original group and look for new groups.Coyote optimization algorithm has the advantage of strong global search ability due to its unique group structure,so it has been widely concerned by scholars at home and abroad.With the in-depth exploration of the coyote optimization algorithm,scholars have found some shortcomings of the algorithm,such as slow convergence speed,low accuracy,easy to fall into local optimum,etc.These shortcomings seriously limit the theoretical development and application scope of the coyote optimization algorithm.This paper improves the coyote optimization algorithm based on its advantages and disadvantages,and applies the improved algorithm to engineering optimization,wireless sensor positioning,UAV path planning and battlefield frequency allocation.The main research work is as follows:(1)An improved coyote optimization algorithm based on dual strategy learning mechanism and adaptive chaotic mutation strategy(DCSCOA)is proposed in this paper.Firstly,it adopts an oscillatory decline factor to generate diverse individuals for enhancing the global search ability.Secondly,it proposes a dual strategy learning mechanism to appropriately increase the influence of the group head wolf,so as to balance the local search ability and global search ability of the algorithm,and to improve the solution accuracy and convergence speed of the algorithm.Finally,it uses an adaptive chaotic mutation mechanism to generate new individuals when the algorithm stagnates,so as to make the algorithm jump out of the local optimum.Through simulation experiments on basic test functions and CEC2017 test functions,the results show that the improved algorithm has higher solution accuracy,faster convergence speed and stronger stability.(2)A multi-objective coyote optimization algorithm(MCOA)based on multi-mechanism fusion is proposed.The algorithm uses the uniform selection mechanism of group head wolves to select group head wolves according to the distribution of the coyote population,so as to improve the distribution of the non-dominant solution set.Then adopt the new growth mechanism of coyotes,use the new growth mode of coyotes and Gaussian mutation to improve the global exploration ability of coyotes.Finally,at the birth and death stage of coyotes,the cross mutation and greedy selection mechanism are adopted to retain the excellent performance of the group head wolves,and then improve the comprehensive performance of the algorithm.MCOA and other comparison algorithms are tested on ZDT and DTLZ test functions.The experimental results show that MCOA has better convergence and distribution than other comparison algorithms.(3)Applying DCSCOA to wireless sensor node localization,a 3D wireless sensor node localization algorithm based on DCSCOA is proposed: DCSCOADV-Hop.The simulation experimental results show that compared with other algorithms,DCSCOADV-Hop can obtain wireless network sensor node localization schemes with better positioning accuracy.A method based on DCSCOA for solving the 3D trajectory planning problem of unmanned aerial vehicle(UAV)is proposed.The simulation experiment results show that compared with other algorithms,DCSCOA can obtain a lower total cost trajectory planning solution.Applying the improved MCOA to the battlefield frequency allocation problem can give a better frequency allocation scheme.