Opportunistic ad hoc networks are traditionally used in remote equipment deployments, sensor networks and VANETs. Applications also include grassroots networks in disaster-recovering areas and privacy-concerned closed-networks. Due to the lack of a model that accurately computes the probability distribution of the delay, we usually content ourselves with the mean values. Such an approach can limit both the ability to predict the system's behavior and the evaluation of the parameters that affect it. In this thesis, we present a framework that allows us to compute analytically the probability distribution of the delay as a function of the eld size, the number of participating users and the movement model. The derived delay probability distribution can help us decide whether opportunistic networking can be practically used in, e.g., dense vehicular environments, highly volatile mesh networks, or in sensor networks helping detect natural disasters in real time. We validate the analytical results against a simulation of the presented model and discuss its accuracy and the parameters that affect it.

The use of the proposed framework can be mostly harvested in dense ad hoc deployments. However, as we increase the density of wireless systems, other problems start to occur, such as high contention and competition for the medium, as well increased number of collisions. Considering the exponential growth of the Internet-of-Things (IoT) applications and the increase in popularity of Vehicle-to-Vehicle (V2V) communications, ad hoc networks have to deal with an ever increasing user density. In order to accommodate the emerging use-case scenarios, the IEEE802.11 protocol is being revised regularly.

However, Carrier Sense Multiple Access/Collision Avoidance (CSMA/CA) has been, for the most part, unchanged throughout the revisions of 802.11 protocol suite. The main purpose of the protocol is to ensure the stability of the network, by avoiding collisions, at the expense of the medium utilization. As ad-hoc, multi-hop, decentralized and distributed networks reemerge to the foreground