| Understanding and reforming the world is eternal topic on human development,although we currently do not know the essence of this world and the universe,a tiny part of them can be seen through theoretical derivation and experimental analysis.Simulation,as an important experimental method,is used to experiment with a model when the experiment cannot be carried out on the real system.Up to now,computer has become the main tool for simulation,while computer hardware resources limit the scale and granularity of simulation,and the scale and granularity of simulation limit the scale and true degree of the research object simulation.With the expansion of the scale of simulation experiments and the increase of granularity requirements of experiments,the traditional self-built high-performance computers or computing clusters have gradually shown their disadvantages in dealing with large-scale simulation applications,including high cost,serious waste of resources,and failure to meet the requirements of heterogeneous and Geographically distributed interconnection and interoperation.Cloud computing and edge computing technology can greatly promote the development of simulation,but it also brings some problems to be solved,like the deployment problem considered in this paper.When simulation is deployed on cloud computing and edge computing resources,simulation applications and resources are no longer coupled.So,it is necessary to consider how to deploy simulation applications on cloud computing and edge computing resources,that is,the matching relationship between various parts of simulation applications and computing resources.Due to the complexity of interaction relationship and resource demand among various parts of simulation applications and network connection relationship and available resource amount between each computing node in cloud and edge computing resources,it is necessary to study how to optimize the deployment of simulation applications under the condition of cloud-edge collaboration to optimize the simulation performance.This problem can be reduced to a combinatorial optimization problem,and the practice has proved that particle swarm optimization algorithm can effectively solve many nonlinear complex optimization problems,so this paper mainly studies the optimization deployment of simulation components in the cloud side collaborative environment based on particle swarm optimization algorithm.The main research work is as follows:1.In this paper,the standard particle swarm optimization(PSO)algorithm is summarized,and the concept of particle exploring space in particle algorithm is proposed for the first time.Based on this,the basic principle,algorithm flow and search process of the algorithm are briefly analyzed and discussed.It provides a new and more direct and effective analysis method for the analysis and application of particle swarm optimization algorithm to solve practical problems.2.The application of particle swarm optimization algorithm in the discrete optimization problem was analyzed,and discussed the differences of premise conditions of the particle swarm algorithm between continuous optimization problems and discrete optimization problems.It is proposed that the change of fitness value of solutions in the solution space will directly affect the performance of particle swarm optimization algorithm to solve discrete optimization problems,and the distance between solutions in the particle search space is the key factor that affects the change of fitness value of solutions in the solution space.At the same time,according to the difference of distance measurement between the solutions in search space,the existing discrete particle swarm optimization algorithm is analyzed and summarized.This analysis is a novel and effective analysis method for the application of particle swarm optimization algorithm in the field of discrete optimization,and provides a theoretical basis for solving discrete optimization problems by using particle swarm optimization algorithm,including the design of discrete particle swarm optimization algorithm to solve the deployment optimization problem in this paper.3.The modeling method of optimal deployment of operational support environment in service-oriented simulation system under cloud computing and edge computing environment is presented.According to the characteristics of deployment problem,a Hamming distance-based discrete particle swarm optimization algorithm(HDPSO)is designed to solve the problem.The HDPSO algorithm is compared with the existing deployment algorithm based on particle swarm optimization and multi-objective genetic algorithm NSGA-II.Simulation results show that the discrete particle swarm optimization algorithm based on Hamming distance proposed in this paper has obvious advantages over other algorithms and can effectively realize the optimal deployment of service-oriented simulation support environment.4.Aiming at the complexity of dynamic user joining and exiting the whole simulation system,the process of user joining and exiting the simulation system and applying for cloud computing and edge computing resources was modeled.The problem was modeled as a deployment optimization problem with uncertain deployment scale,and the HDPSO algorithm was used to solve the problem.In the experiment,The HDPSO algorithm is compared with NSGA-II,and the results show that the HDPSO algorithm still has obvious advantages under the condition of dynamic change of deployment scale,which verifies the effectiveness of the algorithm in solving the deployment problem of user simulation system.. |
