Improvement And Application Of Sine Cosine Algorithm

The sine cosine algorithm is a meta-heuristic algorithm based on the trigonometric function model to solve the global optimal value problem.The algorithm has the advantages of simple structure,few parameters and easy to understand.However,there are also some shortcomings such as slow convergence speed and easy to fall into local optimality,so this paper studies and analyzes the structure and parameters of sine cosine algorithm,makes some improvements to its existing shortcomings,and expands its application field to a certain extent.The research contents of this paper include:(1)A quaternion's coding sine cosine algorithm is proposed.By introducing quaternion coding into the basic sine cosine algorithm,each individual is expressed by four variables,so as to increase the searching information of individuals and the diversity of the population,and further improve the optimization ability and stability of the algorithm.The improved algorithm is simulated by using 19 standard test functions.The results show that the algorithm has good global optimization capability.(2)An adaptive sine cosine algorithm based on behavior evaluation is designed.In this algorithm,the behavior evaluation mechanism of artificial fish swarm algorithm is used to update the individuals,and the sine and cosine updating formulas can update the individuals independently,so as to accelerate the algorithm's search for the global optimal value.The proposed algorithm is tested experimentally with 23 classical standard optimization test functions and an engineering example,and compared with the simulation results of 5 heuristic algorithms,the results show that the adaptive sine cosine algorithm has great advantages in convergence accuracy,convergence speed and algorithm stability.(3)In order to solve the problem of traditional pattern matching with high computational complexity and high computational complexity,an adaptive sinecosine algorithm based on the side inhibition mechanism in the vision system is proposed.Experimental results show that the algorithm can find the target template more quickly and accurately.