Research On Multi-strategy Mayfly Algorithm And Application

Mayfly Algorithm(MA)is a meta-heuristic algorithm that simulates the flight and mating behavior of adult mayflies.It has the characteristics of high computing accuracy and strong search ability,and has been applied in many optimization fields and achieved great success.However,it is found that MA has many shortcomings in solving complex problems,such as too many control parameters and slow convergence speed.In this paper,three multi-strategy mayfly algorithms are proposed and applied to spherical minimum spanning tree,spherical asymmetric traveling salesman and wireless sensor coverage optimization problems respectively to improve the performance of MA and further expand its application range.The main study contents are as below:(1)Because mayfly algorithm has many parameters,the parameter values cannot be determined well,a mayfly algorithm based on the backbone strategy is proposed,which simplifies the basic MA,so that the algorithm does not need parameter control,but determines the position of the mayfly according to the Gaussian distribution and Levy flight.And it uses a "mirror wall" cross-border processing mechanism to reduce the overhead of searching space.Finally,the Bare Bones Mayfly Algorithm is applied to solve the spherical minimum spanning tree problem.The results reveal that the improved algorithm has significant advantages in large-scale spherical minimum spanning tree problem.(2)A Discrete Mayfly Algorithm is proposed.The algorithm uses inver-over operator at the stage of updating mayfly position to increase population diversity and balance the exploration and development capabilities of MA.The crossover operator is used in the mayfly mating stage to accelerate the convergence speed of the algorithm.And 3-opt mutation operation is used for the offspring to get rid of being trapped in local optimum.Finally,the Discrete Mayfly Algorithm is applied to solve the spherical asymmetric traveling salesman problem,and the results show that it can find the shortest path efficiently.(3)To further expand the search scope and exploration capability of the algorithm,an Elite Opposition-Based Bare Bone Mayfly Algorithm is proposed based on the Bare Bones Mayfly Algorithm and the Elite Opposition-based Learning strategy.The algorithm is applied to solve the wireless sensor coverage optimization problem.The experimental results show that the improved algorithm can cover the wireless sensor monitoring area well,while maintaining low redundancy and less energy consumption.