Skip to Main content Skip to Navigation

Offloading strategies for mobile terminals with energy harvesting capabilities

Abstract : Nowadays, the wireless mobile communications are witnessing unprecedented growth fueled by the huge number of connected devices increasing importantly the demands for high-volume data traffic, requiring thus intensive computation and leading to high energy consumption. However, this expansion of wireless services is still restrained by mobile terminals limitations, in terms of processing capacity, storage and energy. Mobile Edge Computing (MEC) and Energy Harvesting (EH) schemes have been recently proposed as promising technologies to extend the battery lives of mobile devices and improve their computing capabilities. On one hand, MEC enables offloading computation tasks from mobile devices to nearby Base Station with more energy and computations resources. On the other hand, EH exploits alternative renewable energy sources to power mobile devices. However, the stochastic nature of renewable energy may lead to energy outage. In such cases, the system’s performance can be degraded due to packet loss or intolerable latency. In order to sensure the system sustainability, efficient transmission policies under EH constraints are needed. In this thesis, we investigate the joint resource scheduling and computation offloading in a single user MEC system operating with EH based devices. The main contribution of this work is the introduction of the strict delay constraint instead of the average delay constraint to satisfy future requirements of lowlatency communications and critical applications. We study three different setups. In the first setup, we consider a perfect Channel State Information (CSI) at the transmitting device and we aim to minimize the packet loss due to delay violation and buffer overflow at the device’s data buffer. The associated optimization problem is modeled as Markov Decision Process and the optimal policy is exhibited through Dynamic Programming techniques. We show that the optimal policy outperforms other policies by adapting the number of processed packets to the system states. In the second setup, we consider a more realistic scenario, where the channel is not perfectly known at the transmitter and it is acquired after an estimation phase. In fact, this estimation can be erroneous degrading thus further the packet loss rate. Hence, we evaluate the previously obtained optimal policy under imperfect CSI conditions and we show that it remains robust with respect to other policies. Finally, we address the setup with no CSI at the transmitter. We therefore assume that an outdated CSI is only available and we show that the proposed optimal policy can still achieve good performance compared to other policies.
Complete list of metadata

Cited literature [106 references]  Display  Hide  Download
Contributor : ABES STAR :  Contact
Submitted on : Tuesday, February 25, 2020 - 12:55:07 PM
Last modification on : Wednesday, June 15, 2022 - 9:03:56 PM
Long-term archiving on: : Tuesday, May 26, 2020 - 3:22:21 PM


Version validated by the jury (STAR)


  • HAL Id : tel-02490626, version 1


Ibrahim Fawaz. Offloading strategies for mobile terminals with energy harvesting capabilities. Réseaux et télécommunications [cs.NI]. Université Paris Saclay (COmUE), 2019. Français. ⟨NNT : 2019SACLT040⟩. ⟨tel-02490626⟩



Record views


Files downloads