ARBR: Adaptive reinforcement-based routing for DTN

Dynamic dual-reinforcement-learning routing strategies for quality of experience-aware wireless mesh networking

“…In [20], a collaborative learning scheme is proposed to increase network throughput by using positive feedback signals to reinforce the popularity of stable routes, and using negative ones to re-route around congested links. Inspired by [20], [21] presents an approach for improving the routing in delay tolerant networks by taking into account mobility, congestion and buffer occupancy metrics. In the scope of wireless sensor networks, [22] proposes FROMS, which is a machine learning-based multicast routing paradigm that can adapt to different cost metrics (e.g., route length or battery levels).…”

Enabling wireless cooperation in user provided networks

“…Adaptive Reinforcement-Based Routing (ARBR) [Elwhishi et al, 2010] uses cooperative groups of nodes to make forwarding decisions based on a cost function at each contact with another node. The protocol considers node mobility statistics, congestion, and buffer occupancy, which are taken as feedback in the cost function.…”

“…Delay Tolerant Routing Solutions based on MARL QLAODV [Wu et al, 2010] Forwarding-based Uses Q-Learning algorithm to change routes preemptively using the learned information AODV ARBR [Elwhishi et al, 2010] Quota-based Groups of nodes cooperate and make forwarding decisions based on a Reinforcement Learning (RL) cost function…”

“…Multi-agent systems have engineering applications in a variety of domains, such as: robotic teams [Balch and Arkin, 1997], intelligent transportation systems [Mhr et al, 2010], games [Niekum et al, 2011], collaborative decision support systems [Bowling and Veloso, 2002], and resource and network management [Elwhishi et al, 2010]. The methodology is based on a set of algorithms and protocols that enable the design of agents which learn the solutions to non-linear stochastic tasks about which the agent has limited prior knowledge.…”

Contention aware mobility prediction routing for intermittently connected mobile networks

“…To countermeasure the deficiency of the previously reported designs, a number of studies were reported [15], [13], [20], [21], [22]. Nonetheless, they are subject to various limitations due to the simplified assumptions related to the nodal mobility scenarios [12], [23], or the utility function presentation [11].…”