Summary:In this subsection, the surveyed literature fo-cused on various metrics and use-cases. Some employedadvanced mathematical techniques such as game theory to im-prove system performance in terms of spectrum utilization [72]or fairness [22]. Using advanced mathematical techniquessuch as stochastic Lyapunov optimization [89] and dynamicprogramming might lead to increased complexity but theyare indeed effective enhancement approaches which give theresearchers insight to evaluation of other metrics such asqueue stability and packet transfer time. The authors of [5]provide the first protocol for signaling and other functionalityin D2D-enabled networks. This helps greatly the researchersand engineers who plan to implement D2D in real world.Nevertheless, the evaluation scenario in [5] can be enhancedto a more realistic setup. Other papers focus on exploiting newuse-cases of D2D communication, such as multicast [9] andcontent distribution [16]. Although most of these papers havenot used advanced mathematical tools, their proposals leadto high performance gains. Moreover, D2D communicationappears to be a viable candidate for applications such asproximity peer-to-peer gaming and social networking.Finally, a summary of the works on underlay D2D commu-nication in cellular networks is provided in Table I, in terms ofmetrics, use-cases, analytical tools, evaluation method,scope,and achieved performances.