The number of Internet of Things (IoT) devices and applications is growing rapidly, as well as the amount of data generated by these devices. To afford various user requirements in data transmission, and reduce bottlenecks, multi-interface intelligent radio devices have increasingly been deployed to provide the flexibility of simultaneous access to several heterogeneous access networks. In this article, we propose a mechanism to dynamically assign the flows of these devices to the appropriate gateway interfaces while satisfying their QoS, and simultaneously increase the valid data accepted by the gateways of our IoT network.
In this thesis, we formulate the problem of optimal assignment of data flows coming from IoT devices having multiple interfaces to several multi-interface IoT gateways (OFAP-MIG) as a Mixed Non-Integer Programming model (MNIP), and propose a solution based on deep reinforcement learning (DRL) approach to solve this NP-hard problem. The agents deployed in the IoT gateways and the central cloud employ Deep Neural Network (DNN) to find the optimal solution based on a state space, an action space and a reward function. The simulation results shows the proposed solution outperforms the centralized one, which has been presented in prior work, in terms of the total amount of data transferred by the gateways, and the acceptance rate.
| Date | 20 Dec 2021 |
|---|
| Original language | French |
|---|
| Awarding Institution | - École de technologie supérieure
|
|---|
| Supervisor | Kim Khoa Nguyen (Supervisor) |
|---|
Amiriyan, M. (Author),
Nguyen (Supervisor),
20 Dec 2021Student thesis: Master's thesis › Master in Engineering: Engineering