微型电脑应用

边缘智能服务的优化发展主要依靠单一的深度学习策略，使得优化后服务时间缩减率较低、通信能耗较高。因此，提出面向人工智能物联网(AIOT)的边缘智能服务的优化技术。根据用户任务请求的迁移和处理流程，构建边缘智能服务网络模型。引入AIOT理念，结合物联网技术和人工智能技术，建立边缘智能服务联合优化框架。以边缘节点服务覆盖范围为基础，从感知时延、服务迁移代价入手，建立自适应服务放置机制。通过任务迁移决策全局搜索方法，生成边缘智能服务最优迁移决策。实验结果表明：所提边缘智能服务优化技术在服务时间缩减率方面比现有优化技术提升了40%以上，有效提升了边缘智能服务效率，且通信能耗始终维持在40 J以下，可节约服务能耗。

The optimization and development of edge intelligent services mainly relies on a single deep learning strategy, resulting in low service time reduction and high communication energy consumption after optimization. Therefore, an optimization technology of artificial intelligence of Things(AIOT) oriented edge intelligent service is proposed. According to the migration and processing flow of user task requests, an intelligent service network model is constructed. The AIOT concept is introduced, by combining Internet of Things technology and artificial intelligence technology a joint optimization framework for edge intelligent services is established. Based on the service coverage of each node, an adaptive service placement mechanism is established from the perception of delay and service migration cost. Through the global search method of task migration decision, the optimal migration decision of edge intelligent service is generated. The experiment results show that the proposed edge intelligent service optimization technology increases the service time reduction rate by more than 40% compared to existing optimization technologies, it effectively improves the efficiency of edge intelligent services, and maintains communication energy consumption below 40 J, which can save service energy consumption.