Improvement Research And Application Of Beetle Antennae Search Algorithm

Optimization problems exist in all areas of life and are closely related to our daily lives.In recent years,the emergence of heuristic optimization algorithms has enriched the theory of optimization problems.Because of its simple implementation and its advantages of easy expansion and high efficiency,it has become a research hotspot.The beetle antennae search(BAS)algorithm is also a heuristic algorithm,which is inspired by the foraging behavior of the beetle.The BAS algorithm has the advantages of simple optimization mechanism,convenient implementation and small computational complexity.Therefore,it has attracted the attention of many scholars at home and abroad and has been successfully applied in many fields.However,the theory of BAS algorithm itself is not perfect enough,and it has the problems of slow convergence,low solution precision,easy to fall into local optimum,and sensitive to parameter setting.Therefore,it is necessary to conduct more in-depth theoretical research.Based on the original BAS algorithm,this paper makes improvements to its shortcomings,and proposes two new algorithms,which are applied to the actual optimization problem solving.The main research contents are as follows:(1)An algorithm called beetle antennae search algorithm based on Lévy flights and adaptive strategy(LABAS)is proposed.And the main innovations include: The algorithm turns beetle into a population and updates the population with elite individuals' information to improve the convergence rate and stability of the algorithm.At the same time,Lévy flights and scaling factor are introduced into the search process to enhance the algorithm's exploration of the potential region of the global optimum.After that,the adaptive step size strategy is used to avoid complicated adjustment operations.Finally,the generalized opposition-based learning is applied to the initial population and elite individuals,which increases the diversity of the population.Experiments show that the LABAS algorithm has the characteristics of high solution precision,fast convergence and good robustness.(2)In this paper,an algorithm called beetle antennae search algorithm with random delay and elite opposition-based learning(BAS-RDEO)is proposed.The main innovations are: The combination of the beetle population and the elite individuals makes the algorithm more stable.In addition,chaotic mapping and opposition-based learning strategies are added to the process of population initialization,so that the initial beetle population with better uniformity and diversity is obtained.In the process of step size updating,not only the information of the elite individuals is used,but also the random time-delay information added is adaptively adjusted according to the evolutionary state of the beetle population,so that the algorithm avoids premature convergence.Finally,elite opposition-based learning strategy and leader multi-learning strategy are used for elite individual,which improves the global and local optimization performance of the algorithm.The experimental results show that the BAS-RDEO algorithm has the characteristics of high optimization precision,fast convergence and good stability.(3)A series of simulation experiments are designed to verify the performance of the proposed LABAS algorithm and BAS-RDEO algorithm.Firstly,the proposed improved algorithms and other representative algorithms are numerically simulated on the standard test functions to compare the optimization performance.In addition,statistical analysis,average convergence curve plotting,and box plot analysis are performed on the experimental results,so that the optimization ability of the algorithms can be comprehensively analyzed.Then,the improved algorithms are applied to the design of tension/compression spring design,welded beam design,and pre-oxidation process control in carbon fiber production with large time-delay characteristics.The experimental results show that the proposed improved algorithms can achieve better optimization results in these practical applications.Therefore,the research in this paper can provide an effective reference for various optimization problems in engineering and industrial production.