: A large number of heterogeneous and mobile devices interacting with each other, leading to the execution of tasks with little human interference, characterizes the Internet of Things (IoT) ecosystem. This interaction typically occurs in a service-oriented manner facilitated by an IoT middleware. The service provision paradigm in the IoT dynamic environment requires a negotiation process to resolve Quality of Service (QoS) contentions between heterogeneous devices with conflicting preferences. This paper proposes a negotiation model that allows negotiating agents to dynamically adapt their strategies using a model-based reinforcement learning as the QoS preferences evolve and the negotiation resources changes due to the changes in the physical world. We use a simulated environment to illustrate the improvements that our proposed negotiation model brings to the QoS negotiation process in a dynamic IoT environment.
Paper
Full text
A Reinforcement Learning QoS Negotiation Model for IoT Middleware
Semantic Scholar · Computer Science · 2020
Abstract
: A large number of heterogeneous and mobile devices interacting with each other, leading to the execution of tasks with little human interference, characterizes the Internet of Things (IoT) ecosystem. This interaction typically occurs in a service-oriented manner facilitated by an IoT middleware. The service provision paradigm in the IoT dynamic environment requires a negotiation process to resolve Quality of Service (QoS) contentions between heterogeneous devices with conflicting preferences. This paper proposes a negotiation model that allows negotiating agents to dynamically adapt their strategies using a model-based reinforcement learning as the QoS preferences evolve and the negotiation resources changes due to the changes in the physical world. We use a simulated environment to illustrate the improvements that our proposed negotiation model brings to the QoS negotiation process in a dynamic IoT environment.