SCIENCE&TECHNOLOGYINFORMATION科技资讯2023NO.16信息与智能科技资讯SCIENCE&TECHNOLOGYINFORMATION基于兴趣度的云VR资源存储方法黄怀龙邹志文(江苏大学江苏镇江212013)摘要:针对云上虚拟现实(VR)资源存储过程混乱和部分云存储空间使用不均衡等存储不当问题,提出一种基于兴趣度的云VR资源存储方法。该方法从虚拟现实资源本身关联度着手分析,定义了兴趣度指数,用于量化云VR资源的关联程度。通过兴趣度指数来构建资源放置组,有效减少了存储过程中的移动次数,并利用兴趣度指数改进的深度强化学习算法进行合理的对象存储设备选择,有效提高了云空间整体使用率。实验结果表明:该方法有效地优化了云VR资源存储过程,解决了资源存储不当的问题。关键词:虚拟现实兴趣度云存储深度强化学习VRUO-CRUSH中图分类号:TP393.01文献标识码:A文章编号:1672-3791(2023)16-0043-08TheMethodforCloudVRResourceStorageBasedontheInterestMeasureHUANGHuailongZOUZhiwen(JiangsuUniversity,Zhenjiang,JiangsuProvince,212013China)Abstract:Aimingattheproblemsofimproperstoragesuchastheconfusingvirtualreality(VR)resourcestorageprocessonthecloudandunevenuseofsomecloudstoragespace,amethodforcloudVRresourcestoragebasedontheinterestmeasureisproposed.Thismethodstartswiththeanalysisoftherelevancyofvirtualrealityresourcesthemselves,anddefinestheinterestindex,andusesittoquantifytherelevanceofcloudVRresources.Theinterestindexisusedtobuildresourceplacementgroupstoeffectivelyreducethenumberofmovesinthestorageprocess,andthedeepreinforcementlearningalgorithmimprovedbytheinterestindexisusedtoselectreasonableobjectstoragedevices,whicheffectivelyimprovestheoverallutilizationofcloudspace.TheexperimentalresultsshowthatthemethodeffectivelyoptimizesthecloudVRresourcestorageprocessandsolvestheproblemofimproperre‐sourcestorage.KeyWords:Virtualreality;Interestmeasure;Cloudstorage;Deepreinforcementlearning;VR;UO-CRUSH传统的虚拟现实技术由于资源所需的硬件成本高昂而发展缓慢,而随着云技术的发展,传统的资源存储方式开始向云存储转变,这将使传统的虚拟现实有了新的发展方向,即云VR[1]。云VR是将传统VR与云技术相结合,利用云进行资源的存储和相关任务的执行。然而由于虚拟现实独特的真实性,其...