UvA/CWI Seminar: Phil Whiting (Macquarie University)
- https://www.cwi.nl/research/groups/scientific-computing/events/uva-cwi-seminar-phil-whiting-macquarie-university
- UvA/CWI Seminar: Phil Whiting (Macquarie University)
- 2018-10-11T12:00:00+02:00
- 2018-10-12T13:00:00+02:00
- Probability and Breaking the Enigma Codes (Part 1) / Measure Valued Mean Field Limits for Random Access Networks (Part 2)
- What Scientific Computing English
- When 11-10-2018 12:00 to 12-10-2018 13:00 (Europe/Amsterdam / UTC200)
- Where CWI, L120
- Contact Name Krzysztof Bisewski
- Web Visit external website
- Add event to calendar iCal
Phil Whiting (Macquarie University)
Part 1: Probability and Breaking the Enigma Codes
Part 2: Measure Valued Mean Field Limits for Random Access Networks
Abstract
Part 1: This talk gives a brief description of the 3 rotor wheel version of the Enigma machine and a probability model for the encoder. It is then explained how breaks could be obtained due to a particular quirk that occurred because of a certain German signalling protocol. These breaks were made well before the start of the Turing-Welchman bombe which began in late summer 1940.
Part 2: Random-access algorithms such as the CSMA protocol provide a popular mechanism for distributed medium access control in wireless networks. We investigate the asymptotic performance of such networks in unsaturated buffer scenarios. These arise in an Internet of Things (IoT) context with highly intermittent traffic sources.
Our results are obtained as a spatial mean field limit, which serves as an approximation to many node random access networks. This limit emerges as a (vector) time dependent measure, and the Partial Differential Equation (PDE) for the underlying densities is presented. As we show, the mean field limit emerges from a sequence of "cluster'' approximations to the interference graph of the underlying prelimit network. Subsequently we develop numerical results for the special case of nodes uniformly distributed on a circle. We conclude with some remarks as to
generalisations of the above construction which might be applied in modelling such networks with additional controls and under more general
assumptions than those presented here.