Research On Multi-strategy Of Beetle Antennae Algorithm And Application

Beetle Antennae Search Algorithm(BAS)is a meta-heuristic optimization algorithm that simulates the beetle’s foraging behavior in natural environment based on their antennae to catch food source odor.The algorithm has the characteristics of fast running speed,simple principle,easy implementation and high efficiency of searching without gradient information.Since the BAS algorithm was proposed,it has been widely used in engineering,economy,energy,planning decision and other fields.With the continuous deepening of study,researchers have explored that the algorithm has many deficiencies,such as the algorithm is easy to fall into local region to get the local optimal solution,it has good solving ability only for lowdimensional problems,and low convergence accuracy of the algorithm becomes weak in the later iteration.In this thesis,we analyze the shortcomings of the BAS algorithm,design the corresponding multi-strategy improvement methods mechanism to enhance the performance of the algorithm and apply the improved algorithm to different optimization problems.My research content expands the theoretical research and application scope of the BAS algorithm.The main work of this thesis are as follows:(1)According to the deficiency that the beetle antennae search algorithm(BAS)is easy to fall into local extremum.We proposed an improved beetle antennae search algorithm(GRBAS)combining the golden ratio in the optimization method with inertia weight.On the basis of analyzing the characteristics of BAS algorithm,the golden ratio method is combined with the length between the two antennae of beetle,and the step length of beetle is updated by adaptive operator,so as to improve the solution ability of the algorithm.We use the benchmark function to test the performance of the GRBAS algorithm,and the excellent performance of GRBAS is compared and verified.The proposed GRBAS is used to solve mathematical problems.It mainly includes solving equations and mathematical numerical integration problems to prove the effectiveness of the algorithm.(2)To obtain better solution performance in different branch problems of path decision planning.An improved algorithm for the optimization of beetle swarm(IBSO)was proposed.The following important strategies are mainly introduced:first,the individual beetle is converted into group beetles for search;Secondly,Levy flight strategy with nonlinear sinusoidal disturbance is introduced to the position of the beetle to enhance the global search capability;Finally,I simulate the characteristics of employing bees in the artificial bee colony algorithm(ABC)to search for better solutions near the honey source field,so as to enrich the solution results of the beetle swarm.The improved algorithm is applied to different problems of path planning,including the traveling salesman problem(TSP)and multi traveling salesman problem(MTSP)in the ergodic optimal path problem in the discrete domain and the two-dimensional and three-dimensional obstacle avoidance problem in the global path planning problem in the continuous domain.The experimental results show that the algorithm has good applicability and provides a new solution for several optimization problems in path planning.(3)To solve the single objective optimization problem in industrial process processing,an Enhanced Beef Antenna Optimization Algorithm(EBSO)was proposed.In order to overcome the limitations of individual beetle antennae search,the individual beetle is converted into a group of beetles,and the mechanism of balancing the direction and allowing the beetle to spiral flight is used to change the straight-line flight of the beetle along its antennae direction.The performance of the EBSO algorithm is evaluated by solving the benchmark function.Finally,we used the EBSO to solve the dynamic optimization in chemical process control reactions,and adopt a new method of discretization of non-fixed points,which is different from the previous method of time domain segmentation.The experimental results show that the EBSO has excellent search performance and achieves satisfactory control trajectory change.(4)In order to solve the contradiction relationship between steady state error and convergence of adaptive filtering in signal denoising.In this thesis,a new method of grey wolf optimization(GWO)based on the hybrid BAS algorithm is proposed to solve the signal denoising algorithm of the normalized variable step size LMS filter.To overcome the disadvantage of incomplete coverage of variables caused by random initialization of gray wolf population,Latin hypercube sampling is proposed to improve the global search ability.Combined with the algorithm of BAS,the leader wolf in the gray wolf group is equiped the "visual" function to improve the local search ability.In order to reconcile the performance between global and local search,we use a nonlinear adjustment factor to adjust each location update of the gray wolf.The MBGWO algorithm is applied to solve the step factor of variable step LMS algorithm in adaptive filter,and the signal is reconstructed according to the determined optimal parameters.In different forms of radar communication signal noise reduction processing.The results show that the proposed method can effectively optimize the noise in signal processing and reduce the original signal distortion.