Weighted Sum-Throughput Maximization for MIMO Broadcast Channel: Energy Harvesting Under System Imperfection

TL;DR

This paper investigates weighted sum-throughput maximization in a MIMO broadcast channel with energy harvesting, system imperfections, and hybrid energy storage, proposing offline and online optimization strategies for different circuit power consumption scenarios.

Contribution

It introduces a comprehensive optimization framework for MIMO broadcast channels with energy harvesting and system imperfections, including novel online policies and bounds for circuit power consumption effects.

Findings

Close-to-optimal online policy for maximum circuit power consumption

Performance bounds established for ideal and maximum circuit power scenarios

Proposed optimization solutions improve throughput under energy constraints

Abstract

In this work, a MIMO broadcast channel under the energy harvesting (EH) constraint and the peak power constraint is investigated. The transmitter is equipped with a hybrid energy storage system consisting of a perfect super capacitor (SC) and an inefficient battery, where both elements have limited energy storage capacities. In addition, the effect of data processing circuit power consumption is also addressed. To be specific, two extreme cases are studied here, where the first assumes ideal/zero circuit power consumption and the second considers a positive constant circuit power consumption where the circuit is always operating at its highest power level. The performance of these two extreme cases hence serve as the upper bound and the lower bound of the system performance in practice, respectively. In this setting, the offline scheduling with ideal and maximum circuit power consumptions are investigated. The associated optimization problems are formulated and solved in terms of weighted throughput optimization. Further, we extend to a general circuit power consumption model. To complement this work, some intuitive online policies are presented for all cases. Interestingly, for the case with maximum circuit power consumption, a close-to-optimal online policy is presented and its performance is shown to be comparable to its offline counterpart in the numerical results.