# Joint Spectrum Allocation and Structure Optimization in Green Powered Heterogeneous Cognitive Radio Networks

## TL;DR

This paper develops a resource allocation framework for OFDM-based cognitive radio networks with energy harvesting, optimizing spectrum and time slot structure to maximize secondary user throughput while ensuring rate requirements.

## Contribution

It introduces a decomposition approach to solve a complex joint resource allocation and structure optimization problem with near-optimal solutions.

## Key findings

- Proposed a sub-channel allocation scheme satisfying rate requirements.
- Reduced the complex problem to a convex optimization task.
- Validated the schemes' optimality through simulations.

## Abstract

We aim at maximizing the sum rate of secondary users (SUs) in OFDM-based Heterogeneous Cognitive Radio (CR) Networks using RF energy harvesting. Assuming SUs operate in a time switching fashion, each time slot is partitioned into three non-overlapping parts devoted for energy harvesting, spectrum sensing and data transmission. The general problem of joint resource allocation and structure optimization is formulated as a Mixed Integer Nonlinear Programming task which is NP-hard and intractable. Thus, we propose to tackle it by decomposing it into two subproblems. We first propose a sub-channel allocation scheme to approximately satisfy SUs' rate requirements and remove the integer constraints. For the second step, we prove that the general optimization problem is reduced to a convex optimization task. Considering the trade-off among fractions of each time slot, we focus on optimizing the time slot structures of SUs that maximize the total throughput while guaranteeing the rate requirements of both real-time and non-real-time SUs. Since the reduced optimization problem does not have a simple closed-form solution, we thus propose a near optimal closed-form solution by utilizing Lambert-W function. We also exploit iterative gradient method based on Lagrangian dual decomposition to achieve near optimal solutions. Simulation results are presented to validate the optimality of the proposed schemes.