Multi-strategy Dynamic Multi-objective Optimization Algorithm Based On Linear Environmental Detection And Its Application

In the fields of industry,economy or scientific research,there are many dynamic multiobjective optimization problems,which not only contain many conflicting objective functions,but also have objective functions changing with time.Because of the time-dependent characteristics of dynamic multi-objective optimization problems,it is not only required that dynamic multi-objective optimization algorithms search for the optimal solution set quickly in the stable time of environment,but also that the algorithm can track the optimal solution set efficiently arrived in the new environment when the changes have occured.Therefore,it is very difficult for dynamic multi-objective optimization algorithms to solve dynamic multi-objective optimization problems.On the basis of multi-objective optimization algorithm,more and more academic researchers embedded the environmental change detection mechanism and environmental change response mechanism,so as to combine them and design dynamic multiobjective optimization algorithm.By performing research and analysis of the existing dynamic multi-objective optimization algorithms in this thesis,we found that the main flow frameworks of the existing dynamic multiobjective optimization algorithms are similar,but most of the dynamic multi-objective optimization algorithms have a single function of environmental change detection mechanism.The environmental change response mechanism is not flexible when the nature of environmental change is not in good matching with the prediction model,which leads to the performance of the optimization algorithms are not satisfactory.In this paper,a new environmental change detection mechanism is designed,which not only detects whether the environment has changed,but also feeds back the attribute information of the environmental change.Multi-strategy multi-objective optimization algorithm based on linear environment detection is proposed.The main work of this paper includes two aspects:Firstly,in view of the shortcomings of the existing dynamic multi-objective optimization algorithms,MSDA-LEM is proposed.The core of the algorithm is to judge the nature of the environmental change according to the migration of population before and after the dynamic change of the environment.According to the different nature of the environmental change,three environmental change response mechanisms are selected to initialize the population,so that the algorithm population has better convergence and diversity in the new environment.Secondly,this paper studies and analyses the background of resource scheduling management in cloud computing environment,and proposes a mathematical model of resource scheduling management in virtual dynamic cloud environment via virtual machine dynamic migration.The model takes into account the total energy consumption of cloud environment and the quality of service about the cloud environment.Based on MSDA-LEM,this paper modifies some of the algorithm operations into MSDA-LEM/Cloud,and MSDA-LEM/Cloud is applied to optimize the model of resource scheduling management in virtual dynamic cloud environment.