Research On Computational Resource Adaptation And Scheduling Method For Open Geographic Simulation

With the continuous development of geographic research in quantitative analysis and systematic synthesis,geographic modeling and simulation has become an important research tool for historical inversion,current situation portrayal,and future prediction of the geographical environment.While geographic modeling research in various domains is in progress,the solution of complex geo-problems has raised the demand for cross-domain model integration applications.Open geographic simulation is a modeling research approach that uses collaborative researcher exploration as the basic pathway.An integrated geographic environment simulation analysis is carried out by organically integrating models,data and computational resources through open geographic simulation,guided by objective geographic laws.With the rapid development of network communication technology and Internet technology,open geographic simulation has put forward a richer demand for computational resources in terms of the evolution of model resources and the updating iteration of data resources.However,existing research related to open geographic simulation has focused mainly on the construction of integrated modeling frameworks and processes,and little has been done on computational resource support methods.Geographic modelling as an effective way to solve complex geographic problems and provide related decision support,the success or failure of a geographic simulation task depends heavily on the support of reasonable computational resources.Therefore,how to provide a stable set of methods to support computational resources in the process of open geographic simulation,is an important element of conducting geographic modeling and simulation work,and also the main starting point of this study.Based on the demand of computing resources for open geographic simulation,this paper investigates a framework for the dynamic aggregation of computational resources that support open geographic simulation.Based on this,this paper explores the computational resource adaptation method for model service operation and the computational resource scheduling method for model service integration.A system of computational resource adaptation and scheduling methods for different simulation scenarios has been developed,so as to support simulation of geographical environment in cyberspace more efficiently.The main research contents and results are structured as follows:(1)The construction of the basic framework of dynamic aggregation of computing resources driven by simulation scenarios.This paper summarizes the connotations and scenario featrues of open geographic simulation,and analyzes the supporting demand of computing resources for open geographic simulation.On this basic,considering modeling researchers,resource providers,model resources,data resources,and computational resources,a basic framework for dynamic aggregation of computational resources driven by simulation scenarios is proposed.Through the compatibility of different network structures and communication modes,we associate the execution requirements of the model with the computing resources,so as to find the optimal computing strategy in the network environment,and further elaborate the basic supporting role of the adaptation and scheduling of computing resources for the open geographic simulation research.(2)The method of computing resource adaptation for single-model service in the network environment is explored.This paper analyzes the key issues of computing resource adaptation for single-model running in the network environment,and proposes an infrastructure for computing resource adaptation.Under the guidance of the infrastructure of computing resource adaptation,this paper designs the description method of running environment information of geographic model and the standardized expression method of computing resource's environment information;this paper proposes an indicator system for matching computing resources and a merit-based recommendation for computing resources to reduce the difficulty of matching model resource and computing resource.Based on the above basic research methods,an intelligent recommended computing resource adaptation method is designed from three aspects: adaptation result recommendation,participatory deployment and model service verification.This method realizes the whole adaptation process from deployment to operation,and provides resource support for the subsequent model integration.(3)The method of computing resource scheduling for multi-model integrated in the network environment is explored.This paper proposes an integrated process representation method based on scientific workflow,which realizes a figurative representation and structured storage of integrated task process.At the same time,a Qo S-based computing resource scheduling strategy is designed to realize the optimizing combination and overall operation control of services in the integration process.Based on the implementation of scheduling strategy,this paper analyzes the optimization scenario in the process of simulating task operation.From the perspective of computing resources,collaborative alternative task optimization strategy and load dynamic aware task scheduling strategy are constructed to guarantee and optimize the operation of integrated task.Aiming at the problem of computing resource demand in the process of open geographic simulation,this paper designs a method of computing resource adaptation and scheduling for open geographic simulation from the aspects of single model operation and multi model integration.On this basis,the system of computing resource adaptation and scheduling for different simulation scenarios is realized to better support the solution of complex geographic problems in cyberspace.