Intelligent Optimization Method And Its Application In Molecular Docking Prediction

With the expansion of control areas and the rise of knowledge automation,the automation of artificial design becomes more and more important.The automation of drug design has important theoretical and practical significance for drug design.In the context of knowledge automation,molecular docking prediction is an important means of drug design automation.This is a complex NP-hard optimization problem that plays a key role in the design of new drugs.However,there is no universal solution to this problem.Therefore,it is of important theoretical and practical significance to further study the intelligent optimization method and apply it to the prediction of molecular docking.We first introduce the development of intelligent optimization methods and molecular docking prediction problems,and choose Differential Evolution algorithm(DE)and Artificial Bee Colony algorithm(ABC)which are widely used in engineering field.According to the characteristics of each algorithm,an improved method is proposed to improve its performance.By adding the learning mechanism and multivariate equation structure to DE algorithm,the adaptive set of the parameters in the search process is realized,so as to enhance the robustness and adaptability of the algorithm.Aiming at the problem of ABC algorithm,the disturbance parameters dynamically adjusting is introduced to enhance the convergence speed of search.At the same time,population information is introduced into the variation equation to improve the direction of population evolution and enhance the global search ability of the algorithm.By the molecular docking prediction experiment,the validity of the two proposed algorithms to the molecular docking prediction problem is verified.However,there is still a problem that the prediction accuracy is low and it is difficult to effectively solve the high-dimensional protein.In order to solve the above problems,first introduce the simulated annealing hyper-heuristics algorithm structure.According to the problem of molecular docking,a specific underlying heuristic algorithm pool is designed,the research group proposed a suitable simulated annealing hyper-heuristics algorithm for molecular docking.According to the test of 23 complex and the widely used GOLD classical test set verified the validity of the simulated annealing hyperheuristics algorithm.Aiming at the shortcomings of the simulated annealing hyper-heuristics algorithm,the heuristic structure of the upper layer and the underlying are improved.Add mutation and cross-operation at the underlying to improve the heuristic combination,introduced population to the upper layer,and strengthen search for the underlying heuristic space.Through the molecular docking prediction experiment,the effective of the improved simulated annealing hyper-heuristics algorithm is verified.