Abstract: As the key entry facility of the identification resolution system, recursive nodes improve the overall service performance by caching identification information and other methods. With the surge of user parsing request traffic, the limited storage space greatly reduces the request hit rate and increases the user query delay. Therefore, a reasonable caching algorithm needs to be designed to improve the resource response speed. Aiming at the cache problem in the identification resolution scenario, considering the spatial heterogeneity of user content requests and the hierarchical structure of identification data, a cache content deployment scheme based on user content preference prediction is proposed. Firstly, based on the user’s historical request information, the user’s content preference is calculated by using the identifier features, and a method based on polynomial prediction is used to predict the user’s request for the identifier; then, according to the predicted request, the improved genetic algorithm is used to design content deployment scenarios. The simulation results show that the proposed algorithm is superior to the comparison algorithms in terms of cache hit rate and delay performance, and the proposed genetic algorithm has a faster convergence speed.

Key words: industrial Internet, identification resolution, caching strategy, genetic algorithm, dynamic popularity

摘要： 作为标识解析体系的关键入口性设施，递归节点通过缓存标识信息等方法提升整体服务性能。随着用户解析请求流量激增，有限的存储空间使得请求命中率大幅下降，用户查询时延增加，因此需要设计合理的缓存算法以提高资源响应速度。针对标识解析场景中的缓存问题，考虑用户内容请求的空间异构性和标识数据的层次结构性，提出了一种基于用户内容偏好预测的缓存内容部署方案。基于用户的历史请求信息，利用标识的数据特征统计用户的内容偏好，并使用一种基于多项式回归的方法来预测用户对标识信息的请求情况；根据预测的请求情况，利用改进的遗传算法设计内容部署方案。仿真结果表明，所提算法在缓存命中率与时延性能方面均优于对比算法，同时所提的改进遗传算法具有更快的收敛速度。

关键词: 工业互联网, 标识解析, 缓存策略, 遗传算法, 动态流行度