Performance analysis of online matching algorithms for D2D communications

Abstract

In this paper, we consider a Device-to-Device (D2D) cellular network in which idle users can work as relays between cell users and the base station to improve their data rate. The relaying induces a cost for the User Equipment Relays (UER), that should be compensated with a payment from the mobile operator so that UERs accept to offer the service. The problem hence arises for the operator to match cell users and UERs at a reasonable cost and increasing the data rate. In this context, we consider the requirements of truthfulness, budget feasibility and acceptance of online scenarios to compare ON algorithm, which considers all constraints, with other three algorithms that were not built to respect all of them, Hungarian, Threshold and Online Weighted Knapsack (OWK). We observed that ON algorithm is the best in terms of execution time; however, it does not scale well considering the number of matched edges, requiring modifications in its selection criteria. In addition, we noticed that OWK algorithm has appealing properties and, if it were modified to be truthful and to reduce its complexity, it would present the best results.