Design of Space-ground Integrated Simulation Software Based on Algorithm Pluginization
In recent years,with the rapid development of integrated space-ground networks,simulation demands have gradually shifted towards complex scenarios such as large-scale constellation deployment and multi-user concurrency.To address the performance bottlenecks of integrated space-ground simulation software in complex scenarios and meet the needs for continuous algorithm iteration and upgrading,we propose a design for integrated space-ground simulation software based on algorithm pluginization.Firstly,we design plugin connectors,standardize algorithm model access specifications,and unify interface communication protocols to ensure effective integration of algorithms of different types and programming languages with the simulation platform.Secondly,we adopt a cloud-native architecture,packaging algorithm plugins into containers and using container orchestration technologies to deploy these plugins in a distributed system,thereby simplifying the deployment process and supporting dynamic scaling of plugin replicas.Lastly,we design a plugin computation framework that supports task splitting,allocation,and result merging.The allocation is achieved by load balancing and scheduling sub-tasks to multiple plugin replicas for parallel computation.Experimental results demonstrate that the simulation software based on algorithm pluginization significantly enhances system computational performance and meets the simulation needs of complex integrated space-ground networks.
万方数据
维普
