Multi-strategy Optimization Research On Artificial Bee Colony Algorithm

In recent years,researchers have proposed many intelligent computing models to solve practical engineering problems based on various biological behaviors and phenomena in nature.The Artificial Bee Colony algorithm(ABC),as a kind of intelligent optimization algorithm,has become a hotspot in recent years due to its characteristics of few control parameters,easy to implement and simple structure.This topic is based on understanding and researching about the Artificial Bee Colony algorithm.In view of some shortcomings such as the global and local searching ability is not balanced and easy to fall into local optimal value in the late,it puts forward two kinds of different improved artificial colony algorithm---an Artificial Bee Colony algorithm with Gaussian distribution scaling factor(GDABC)and an Artificial Bee Colony algorithm with Intraspecific Competition on computing resources(IC ABC).The main work is as follows:1.An artificial bee colony algorithm with Gaussian distribution scaling factor is proposed to strengthen exploration on the basis of enhancing the exploitation.A searching strategy which learns from the optimal solution based on neighborhood is used in the employed bee phase which could increase the diversity of population.A searching equation based on the trigonometric mutation which could adjust searching step adaptively and modify the way of choosing foods source.The scaling factor of two searching equations obey Gaussian distribution.It can realize the balance between exploration and exploitation through using the characteristic of two equations and setting Gaussian parameters by experiment.Finally,28 standard test functions with different characteristics are used in simulation experiments to verify the search quality,convergence speed and robustness of the improved GDABC algorithm.2.In order to solve the disadvantages of the weak role of the roulette stage in the original ABC algorithm,a competitive computing resource allocation mechanism 1s proposed.In the leading and onlooker bee stages,multiple search strategies are used to simulate the intra-species competitive environment in the way of inclined computing resource allocation.The scout bee uses its structural characteristics to propose an equipotential replacement model.Members in external documents that do not occupy computing resources provide population diversity at the same time and strengthen the development of potential areas while ensuring exploration.Finally,22 benchmark functions are used with 30 functions proposed by CEC2014 special session to analyze and discuss the comprehensive performance of ICABC algorithm.In above two improved algorithms,the one focuses on the balance betvreen exploration and exploitation,and the other nlakes full use of intra-species competition node to improve the optimization efficiency of the algcorithm from the perspective of computational resource allocation.Both of them have high optimization performance and are verified by experimental verification.