Dealing with Dynamic Behavior

Algorithmic research has led to progress in understanding more and more realistic models of wireless networks. By and large, however, these models have been static: they assume that the network has certain functionality which remains the same throughout the duration of the computation. This is perhaps the issue where theory and practice differ the most. Empirical scientist have long observed that wireless networks and their environments are highly dynamic and often unreliable.

Dynamic behavior comes in many flavors. The simplest one is where nodes come and go, possibly due to mobility. More subtle case is where links come and go. More generally, reception signal strength can vary for a multitude of reasons, often due to ordinary changes in the environment (people moving around, doors open, humidity changes).

General purpose algorithmic wireless models should provide a level of robustness to such dynamic behavior. Managing dynamicity by ad-hoc measures is not likely to lead to good algorithms. The challenge faced by the algorithm community is to combine robustness with other essential features of good algorithmic models: generality, analyzability, and computability.

At this year's workshop we will curate a collection of invited talks that will introduce a variety of different models and matching purposes, along with cutting-edge results and open questions in these settings. Plenty of time will be left for discussion and exploration of collaboration opportunities.

Invited presentations

Martin Hoefer, MPI Informatik: No-Regret Learning in Dynamic Wireless Networks

Jaikumar Radhakrishnan, TIFR : Communication lower bounds that are sensitive to distances between processor

Kazunori Hayashi, Kyoto University: Channel estimation for multicarrier systems using compressed sensing

The focus of WRAWN

Much remains to be understood about the algorithmic complexity and efficiency of wireless networks, despite their near omnipresence. The focus of WRAWN is on keeping researchers in this field up to date on cutting edge models, problems, and approaches. Each year's meeting chooses a specific theme relevant to this general goal. In previous years, for example, the workshop has dived deep into understanding advances in SINR-style models, probed the gap between theory results and practical implementations, and examined situation-specific models.