| dc.contributor | 資科系 | |
| dc.creator (作者) | 張宏慶 | |
| dc.creator (作者) | Jang, Hung-Chin | |
| dc.creator (作者) | Chiu, Chr-Jr | |
| dc.date (日期) | 2021-10 | |
| dc.date.accessioned | 7-Oct-2022 14:41:38 (UTC+8) | - |
| dc.date.available | 7-Oct-2022 14:41:38 (UTC+8) | - |
| dc.date.issued (上傳時間) | 7-Oct-2022 14:41:38 (UTC+8) | - |
| dc.identifier.uri (URI) | http://nccur.lib.nccu.edu.tw/handle/140.119/142380 | - |
| dc.description.abstract (摘要) | Smart homes provide users with a more convenient, comfortable, and safe living environment through various automation equipment, high-tech home appliances, and network services. With the development of the Internet of Things and widespread smart home, there are many Internet-enabled services in smart homes. Each kind of service has distinct service quality requirements. Remote disaster warning and remote monitoring and detection service emphasize realtime and reliability. UHD video (4K/8K) and VR /AR services require high transmission bandwidth. As the number of IoT-enabled equipment of smart homes increases, it becomes imperative to effectively allocate limited bandwidth resources and improve the overall network performance to ensure the effectiveness of various services. From the perspective of Internet service providers (ISP), this research studied deep reinforcement learning techniques cooperating with software-defined networks (SDN) to improve traditional smart homes` bandwidth management architecture and bandwidth allocation method. The SDN architecture separates the control plane and the data plane. It centralizes the control mechanism to simplify and support flexible network management. Deep reinforcement learning does not rely on labeled data instead of exploring the unknown environment and the environment`s feedback on current as the basis for the subsequent action. This study used SDN to simulate the network environment from smart homes to an ISP. Due to the lack of a sufficient amount of labeled data, and it is not easy to establish a standard for data labeling, we used the Deep Q-network (DQN) of deep reinforcement learning in the bandwidth allocation for smart homes. The simulation results show that Deep Q-network can achieve high performance in both the jitter reduction ratio and the probability of successfully fulfilling bandwidth allocation termination conditions. | |
| dc.format.extent | 108 bytes | - |
| dc.format.mimetype | text/html | - |
| dc.relation (關聯) | 2021 IEEE 12th Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON), pp. 98-101 | |
| dc.subject (關鍵詞) | Deep Q-Network; Bandwidth Allocation; Smart Home | |
| dc.title (題名) | Using Deep Q-Network in Bandwidth Allocation of Smart Homes | |
| dc.type (資料類型) | conference | |
| dc.identifier.doi (DOI) | 10.1109/IEMCON53756.2021.9623084 | |
| dc.doi.uri (DOI) | https://doi.org/10.1109/IEMCON53756.2021.9623084 | |
