Resource Allocation in Green Dense Cellular Networks: Complexity and Algorithms

TL;DR

This paper addresses the complex problem of resource allocation in energy-harvesting dense cellular networks, characterizing its NP-hardness and proposing efficient algorithms with theoretical and simulation validation.

Contribution

It provides the first comprehensive analysis of the problem's NP-hardness and introduces tailored algorithms for different network scenarios with proven performance bounds.

Findings

Polynomial-time optimal algorithms for single channel, single EBS cases.

An efficient approximation algorithm for multiple EBSs with one channel.

A heuristic algorithm for multiple channels, validated by simulations.

Abstract

This paper studies the problem of user association, scheduling and channel allocation in dense cellular networks with energy harvesting base stations (EBSs). In this problem, the EBSs are powered solely by renewable energy and each user has a request for downloading data of certain size with a deadline constraint. The objective is to maximize the number of associated and scheduled users while allocating the available channels to the users and respecting the energy and deadline constraints. First, the computational complexity of this problem is characterized by studying its NP-hardness in different cases. Next, efficient algorithms are proposed in each case. The case of a single channel and a single EBS is solved using two polynomial-time optimal algorithms---one for arbitrary deadlines and a less-complex one for common deadlines. The case of a single channel and multiple EBSs is solved by proposing an efficient constant-factor approximation algorithm. The case of multiple channels is efficiently solved using a heuristic algorithm. Finally, our theoretical analysis is supplemented by simulation results to illustrate the performance of the proposed algorithms.