Multi-Agent Model Of Large-Scale Geospatial Simulation Based On Parallel Optimization

The unbalanced phenomenon in the history of human social development has caused many irreparable costs,and it is urgent to solve such problems by scientific means.The multi-agent simulation model makes use of its bottom-up advantages,integrates the geospatial environment at the bottom,the behavior rules of agents at the middle level and the agents at the top level,and clearly shows the dynamic feedback process between human society and natural systems.It is one of the best methods to study the "man earth relationship".However,the relationship between the agents in the complex system is complex,and the seemingly unrelated two variables are actually interdependent.This natural law makes the multi-agent simulation model difficult to calculate.With the increasing scale of the simulation space,the operation efficiency of the model makes the research more difficult.Parallel computing focuses on dealing with large-scale complex computing problems.The efficiency of large-scale complex system modeling can be effectively solved by building parallel multi-agent models.This research combines the multi-agent idea and parallel computing technology,adopts the "divide and conquer" principle,designs the parallel and serial hierarchical structure according to the time node interval,and constructs a parallel multi-agent model based on the decomposition mapping agent task scheduling optimization method.In order to verify the feasibility of the model,the grassland ecological policy driven multi-agent model and the planting artificial grass multi-agent model were taken as examples to verify.The simulation results show that the parallel multi-agent model applied to grassland ecology can accurately and efficiently predict the evolution law of artificial grassland and the future grassland and the production and life order of herdsmen under different policies,which is helpful for the government to grasp the decision-making direction and maintain the grassland ecological balance.After the decomposition mapping agent task scheduling optimization,the model results are basically stable,the computing efficiency is significantly improved,and the acceleration ratio is about 3.3.However,with the increase of parallel tasks,threads switch frequently,which has a certain impact on the parallel efficiency.The experimental results show that the parallel method based on decomposition mapping agent task scheduling can be fully compatible with the multi-agent model.It is suitable for all geospatial multi-agent simulation based on the relationship between man and earth.It greatly improves the simulation efficiency of large-scale geospatial models,and has important reference significance for the research of large-scale complex systems.