Algorithm Research On Butterfly Optimization Algorithm Based On Hybrid Strategy

Butterfly Optimization Algorithm(BOA)is a new meta-heuristic optimization method designed to simulate the foraging behavior of butterflies.Because the butterfly optimization algorithm is intuitive,simple and effective,and easy to implement,it has been widely used to solve various complex optimization problems.With the deepening of the research,the researchers found that the algorithm has poor development ability,easy to fall into the local optimum and slow convergence speed in the later stage.Aiming at the shortcomings of the butterfly optimization algorithm,this thesis optimizes the algorithm from two aspects of algorithm hybridization and multi-strategy mixing,and applies the improved algorithm to optimization problems,further perfects the BOA algorithm theory and expands its application range.The main work of this paper is as follows:(1)Propose a hybrid butterfly and flower pollination meta-heuristic algorithm(MBFPA)based on a mutually beneficial symbiosis mechanism.First,the flower pollination algorithm has good exploration capabilities,and the organic blending of the butterfly optimization algorithm and the flower pollination algorithm greatly improves the algorithm’s exploration capabilities;second,the Symbiosis Search(SOS)has strong development capabilities in the mutually beneficial symbiosis stage.By introducing a mutually beneficial symbiosis phase,the development capability of the algorithm is effectively improved,and the convergence speed of the algorithm is accelerated.Finally,the adaptive switching probability is increased,which improves the balance of the algorithm’s detection and development capabilities.In order to evaluate the effectiveness of the algorithm,in 49 standard test functions,the algorithm was compared with 6 basic meta-heuristic algorithms and 5 mixed meta-heuristic algorithms.MBFPA is also used to solve four classic engineering optimization problems.The comparison of experimental results shows that the method is feasible,and MBFPA has strong competitiveness and application prospects.(2)The proposed MBFPA algorithm is applied to the optimized extreme learning machine,and 10 standard UCI data test sets are selected for training.The experimental results show that compared with other meta-heuristic algorithms,the optimized extreme learning machine has better performance.(3)A multi-group interactive butterfly optimization algorithm(MGIBOA)is proposed.The introduction of multiple groups is used to increase the diversity of the population and increase the communication of the population;the introduction of the Levi flight trajectory and the Gaussian random walk alternate strategy greatly improves the algorithm’s convergence speed and local development capabilities.MGIBOA was used for 13 classic benchmark function tests,compared with other 6 well-known heuristic algorithms,and used to solve the IIR pattern recognition problem.The simulation experiment results show that the algorithm has faster convergence speed,higher accuracy and stability.