Research On Resource Scheduling Model Based On Computational Complexity Of Vector Big Data

In the context of the era of big data,geographic computing has begun to face the problems of data quantification and multi-tasking resource scheduling.How to allocate different computing tasks to corresponding resource nodes to run to achieve the goal of optimal resource allocation is a difficult problem facing today.At present,parallel computing is an important method for solving massive data processing.Because the allocation of computing resources will directly affect the performance of parallel tasks,and the impact of computing resources on different parallel tasks varies,it is extremely difficult to schedule resources for parallel tasks.In the study of parallelized spatial analysis algorithms,this paper takes parallelized intersection algorithm as an example.Aiming at the problems of large complex polygon intersection computation time and uneven distribution of data and computing resources,a high-performance algorithm based on computational complexity prediction is proposed.Turn method.First,multi-level bounding boxes are used for secondary filtering to Reduce the one-to-many relationship of the intersecting objects.Then,the method of filtering holes and islands by using bounding boxes is used to Reduce polygons that are not involved in calculation and optimize the processing process of complex polygons.The non-linear regression model measures the computational complexity of the intersection operation and optimizes the parallel computing load accordingly.Experimental results show that the method proposed in this paper effectively improves the intersection efficiency of massive and complex polygons.In the research of resource scheduling,this paper researches parallel spatial analysis tasks from two aspects of data complexity and algorithm complexity,and proposes resource scheduling strategies for parallel spatial analysis tasks.First,based on the historical job information,a parallel algorithm complexity prediction model is established to accurately express the performance change trend of the parallel algorithm;then based on this model,a static resource scheduling model that takes into account resource utilization is established;finally,a dynamic resource scheduling model is established based on the idea of ??genetic algorithms.The experimental results show that the method proposed in this paper can dynamically schedule resources according to the predicted running time of different tasks to achieve the minimum task running time.The experiments prove that this paper solves the resource scheduling problem of multiple parallel tasks based on the idea of genetic algorithm.