A general utility optimization framework for energy-harvesting-based wireless communications

Abstract

In recent years, wireless communication systems are expected to achieve more cost-efficient and sustainable operations by replacing conventional fixed power supplies such as batteries with energy harvesting (EH) devices, which could provide electric energy from renewable energy sources (e.g., solar and wind). Such EH power supplies, however, are random and instable in nature, and as a result impose new challenges on reliable communication design and have triggered substantial research interests in EH based wireless communications. Building upon existing works, in this article, we develop a general optimization framework to maximize the utility of EH wireless communication systems. Our framework encapsulates a variety of design problems, such as throughput maximization and outage probability minimization in single-user and multiuser setups, and provides useful guidelines to the practical design of general EH based communication systems with different assumptions over the knowledge of time-varying wireless channels and EH rates at the transmitters.

Optimal Resource Allocation for Wireless Powered Mobile Edge Computing with Dynamic Task Arrivals

“…, K } of single-antenna users, and provides cloud-like computation capability to enable their computation-intensive applications. We consider a finite time horizon with duration T , which 1 Notice that the offline optimization approach has been widely adopted in the energy harvesting based wireless communication [14] and the energy management in smart grids [15]. The obtained optimal offline solution can serve as the performance upper bound for any practical online designs, and can also inspire practical online designs in the case when the CSI and the task arrival information are casually known.…”

“…B. Obtaining Optimal Dual (λ opt , µ opt , ν opt ) Next, we maximize G(λ, µ, ν) over (λ, µ, ν) to solve the dual problem (14). Note that the dual function G(λ, µ, ν) is always concave but not necessarily differentiable.…”

“…Note that the dual function G(λ, µ, ν) is always concave but not necessarily differentiable. Therefore, problem (14) is convex, and can thus be solved by subgradient based methods such as the ellipsoid method [16], in which the subgradient of G(λ, µ, ν) is…”

“…Therefore, it is an important yet challenging problem to jointly design the WPT at the AP and the computation and communication resource allocation at the users, in order to properly control the users' harvested energy to well match their dynamic computation requirements in an energy-efficient manner. Notice that in the literature, some prior works [11]- [13] investigated joint communication and computation management in MEC systems with dynamic task arrivals over time, while other works (see, e.g., [14], [15] and the references therein) considered dynamic energy management with random energy arrivals for energy harvesting based systems. However, there still lacks a joint design of the WPT at the AP and the communication/computation energy demand at the users for wireless powered MEC systems, by taking into account both energy and task dynamics over time and space.…”

A Low-Complexity Policy for Outage Probability Minimization With an Energy Harvesting Transmitter

“…Similarly, with only CDI at the transmitter, [6] extends [5] to a system with EH transmitter and receiver. In [7], an infinite battery is assumed and a general framework is set for utility maximization in EH communication systems. In [8], outage probability is studied in a delay-limited EH system with finite and discrete energy, channel and battery states.…”