学术报告

当前位置: 网站首页 > 科学研究 > 学术报告 > 正文

第26期学术报告:传感云中基于强化学习博弈的实时映射容量优化研究

时间:2019-06-25 15:19:24  作者:  点击: 107 次

26期学术报告

题 目传感云中基于强化学习博弈的实时映射容量优化研究

时  间2019年6月23日15:30

地  点梁林,第二食堂 328

报告人刘建华

    博士,绍兴文理学院计算机系副教授。研究方向为无线网络与通信、物联网、网络安全、大数据分析。主持国家自然科学基金1项,主持完成浙江省自然科学基金1项。在SCI/EI期刊上发表论文20余篇。

       Abstract: Sensor cloud is an emerging edge node assisted technology that offers a range of flexible configurations for massive sensor radio access from clusters of sensors with heterogeneous requirements. A configuration specifies the amount of radio resource allocated to each cluster of sensors for random access and for data mapping to the cloud platform by edge computing. Assuming no knowledge of the traffic statistics, there exists an important challenge to determine the configuration that maximizes the long-term average number of sensor nodes at each mapping time interval in a real-time fashion. Given the complexity of searching for optimal configuration, we develop real-time configuration selection based on the tabular Q-learning, the linear approximation-based Q-learning, and the deep neural network-based Q-learning (DQN) in the single-parameter single cluster scenario. Our results show that the proposed reinforcement learning-based approaches considerably outperform the conventional heuristic approaches based on load estimation in terms of the number of mapped sensor nodes. We further proposed a cooperative multi-player learning(CMPL) for the multi-parameter multi-cluster scenario, and the CMPL also outperforms the conventional heuristic approaches in both capacity and training efficiency.

 

地址:嘉兴市嘉杭路118号 邮编:314001 E-Mail:jx83640102@163.com
版权所有 Copyright(C)2011 嘉兴学院数理与信息工程学院