Improved Crow Search Algorithm And Its Application Research

The Crow Search Algorithm(CSA)is a novel population-based metaheuristic algorithm proposed by Iranian scholar Alireza Askarzadeh.The algorithm simulates the behavior of crows storing extra food and retrieving it when necessary to establish relevant mathematical models.CSA is characterized by its clear structure,few adjustable parameters,fast convergence speed,and high precision.However,research on the CSA algorithm is still in its early stages,and it has some limitations,such as weak local optimization capabilities and a blind location updating strategy.In this paper,we address these limitations and propose an improved and optimized version of the CSA algorithm,which provides a new method for solving complex problems and extends its use in optimizing various fields.The main contributions of this paper are as follows:(1)To reduce the blind updating of the crow search algorithm and prevent the algorithm from getting trapped in local optima,we use an adaptive adjustment strategy,introduce an experience factor and Levy flight,and propose an adaptive crow search algorithm based on Levy flight(LACSA).We compare LACSA with other novel metaheuristic algorithms on 20 classical benchmark functions,and the results show that LACSA outperforms the other algorithms in function optimization problems.(2)To further balance the algorithm’s exploration and exploitation abilities,we propose a self-adaptive crow search algorithm combined with the Particle Swarm Optimization algorithm(PLACSA).Based on LACSA,the algorithm introduces the information sharing mechanism of Particle Swarm Optimization to give crows a sense of group concept.We enhance the search ability of the optimal individual by using Cauchy mutation mechanisms.We apply the PLACSA algorithm to engineering design and wireless sensor network coverage optimization problems and compare it with other algorithms.The experimental results show that the PLACSA algorithm has strong performance in solving optimization problems.(3)As an application,we use the proposed self-adaptive crow search algorithm combined with the Particle Swarm Optimization algorithm to solve the threedimensional unmanned aerial vehicle(UAV)path planning problem.By solving multi-objective constraint functions,we determine the flight path of the UAV.We compare the improved algorithm with other novel metaheuristic algorithms in different benchmark scenarios.Simulation results show that the algorithm performs better in finding the optimal path that minimizes the cost function compared to other algorithms.