Tim Boxer

Dr Tim Boxer

Programme Manager

Industrial Mathematics KTN

Blogs

TestTB2

this is a test

TestTB

This is a test

Exploring and explaining - effective knowledge networks

A knowledge network can support prioritisation by providing mechanisms that enable people to explore their interests in the context of the community’s knowledge.  Industry knows what challenges it faces, but not necessarily what mathematics might be relevant to their solution.  Eliciting and exploring expressions of interest contribute to the knowledge base that underpins the focus and direction of a vibrant knowledge network.

A knowledge network can also support contextualisation by providing mechanisms that explain the interpretation of project findings and expressions of interest in the context of the shared knowledge base.  This is important for the assimilation of relevant knowledge by the community.  An effective knowledge network provides an environment that stimulates the sharing of relevant knowledge through exploration and explaination. 

 

If you have an industrial challenge to explore: Study Group 2012

If you want GPU computing explained: The GPU computing revolution

Return to main article: What makes an effective knowledge network?

Investment and publishing - effective knowledge networks

A vibrant knowledge network includes investment in new projects within the scope of the network.  Resources for proposing and undertaking activity are limited and suitable mechanisms are required to elicit high quality proposals and to fund the best.  An effective knowledge network enables the community to determine its investment priorities on the basis of a shared understanding of the gaps in knowledge that would be most fruitful to fill.

Publication of project findings (where possible) is an important part of ensuring that the community is well informed about recent progress.  However, a knowledge network is more effective if it has mechanisms to support understanding of new findings in the context of the community’s knowledge.

 

If you wish to propose a short project: IM-sKTP application process

Some examples of recent project findings: Published case studies

Return to main article: What makes an effective knowledge network?

Scope and interaction - effective knowledge networks

A knowledge network can be thought of as a community of people involved in project activity relevant to the scope of the network and interacting to share knowledge.  The relevant knowledge for a network in industrial mathematics includes not only mathematics but also the domain knowledge about challenges from industry and scientific knowledge from other disciplines.  

The community of people includes industry, academia, government, the research councils and the Technology Strategy Board.  A knowledge network is more effective if the scope is chosen well and if fruitful interaction is enabled by mechanisms that provide environments to encourage sharing of relevant knowledge.  

 

Review our project portfolio: Projects listed by company and university

See how mathematics cuts across sectors: Projects listed by sector

Return to main article: What makes an effective knowledge network?

 

Providing the evidence for evidence-based decision-making

On 15 November 2011 the CDE launched a call for for research proposals that will provide new and innovative approaches, methods, tools and techniques for Operational Analysis.

Funding available: Total £500k for this call.  Note, that they are expecting each proposal to be £30-£50k proof of concept and for < 6 months duration. 

Call documents and slides presented at the launch event are available at http://www.science.mod.uk/events/events_and_calls.aspx.

HM Gov is committed to evidence-based decision-making.  Operational Analysis (OA) is used by MoD to provide a structured approach to decision making: bringing together data and expert knowledge in a transparent way.  OA can be used to examine trade-offs between policy aspirations and affordability constraints.

OA provides quantitative understanding of issues underpinning top level decision-making.  There is no room for black box models: there must be an auditable trail.  MoD is always looking for ways in which the responsiveness of their analytical capability can be improved.  

General challenges / requirements are:

  • reducing the time and cost of developing and employing analytical models and tools
  • credible representation of emerging challenges and capabilities
  • combining numerical and judgemental based methods
  • visualising and communicating complex data and conclusions
  • balance of investment across disparate capabilities and lines of development.

The call document provides more information about the areas covered by the call.  Proposers are encouraged to get in touch with CDE before expending significant effort on preparing a proposal.

 

Call closes 17 Jan 2012 (1200 noon)

Stochastic modelling in energy systems

As part of the LANCS Initiative in Operational Research, researchers are building capability in the area of stochastic modelling.  Stochastic modelling is relevant to problems in a wide range of energy systems, from energy production, distribution and transmission to pricing and beyond. An understanding of stochastics is vital for the use of modelling to support decision making for energy systems.

A 2-day workshop on Stochastic Modelling in Energy Systems will be held in Lancaster from 1 pm Monday 30 January 2012 to noon Wednesday 1 February 2012. 

The goal of the workshop is to share existing work and establish new links and projects within this fascinating area.  Energy industry participants are encouraged to attend and are invited to contribute a talk about their challenges and approaches to decision-making under uncertainty.

The workshop will provide an opportunity to hear from distinguished international researchers in this field, in particular:

  • Professor Asgeir Tomasgard from the Norwegian University of Science and Technology in Trondheim will deliver a tutorial on his energy-related work with industry and government.
  • Professor Andres Ramos from Universidad Pontificia Comillas in Spain is a leading energy researcher in Europe, and will present some of his work on stochastics in energy planning.

Participation will be by invitation only, if you are interested, please contact Dr Tim Boxer.  Please also contact me if you are interested in this area, but unable to attend the event.  A detailed programme is now available.

Underpinning Defence Mathematics Event, 24 March 2011

The Ministry of Defence (MoD) is supporting a Programme of Underpinning Defence Mathematics (UDM) through the Industrial Mathematics Knowledge Transfer Network.

The UDM Programme has set up three research contracts and four Internship projects which are now nearing completion.  The programme review seminar on 24 March 2011 in London is an opportunity to hear about the findings of the UDM Programme.  Attendance is by invitation only.  If you wish to express interest in attending, please complete the online form at www.industrialmath.net/udm/event_request.html.  The event will include presentations about the following UDM projects:

 

Research contracts:

  • Efficient entropy-based detection of change-points in streaming data(Bristol)
    This research contract is developing algorithms for detecting change points in a data stream based on match length methods, and on locating approximately mutually singular data segments. The match length algorithm output is used directly rather than simply as a way to estimate entropy. Testing these algorithms on real data and in particular on the output of compressed sensing is expected to provide new insights.
  • Information driven quantisation for image compression(Warwick / Liverpool)
    This research contract is working on new methods of image quantization that are particularly effective at low signal-to-noise ratios. These methods, optimising the bit-depth used while retaining the information content, result in good image compression. The project also covers the registration of successive images with variable bit-depth, and reconstruction algorithms from a sequence of such images.
  • Novel data representations for information estimation(ThinkTank Maths)
    This research contract is developing ways of measuring the information content of a data stream for a particular purpose. Effective methods for rapidly estimating the information content of raw sensor data would reduce both bandwidth requirements and the need to apply computationally expensive algorithms. This project aims to achieve this objective by creating novel data representations and information measures.

 

Internship projects:

  • Clutter mapping for radar trackers(Thales / Bath)
    Radars can produce unwanted detections (clutter) due, for example, to reflection from the ground. It is important to estimate the spatial density of the clutter, so that the probability of outputting false radar tracks can be controlled. The aim of this Internship project is to perform a rigorous analysis of the problem, and for each potential solution, to consider efficient strategies for computation on real systems.
  • Optical correlators(MBDA and Cambridge Correlators / Sheffield)
    A compact optical computer has already been demonstrated to achieve basic image correlation and tracking functions. However, the high speed of processing is such that it could have wider application in computational imaging, where the processing can compensate for the use of simple or sparse sensing to recover imagery at a quality. The aim of this Internship project is to develop real time techniques and explore potential applications in seekers for guided missiles.
  • Topology of antenna radiation patterns(Selex / Edinburgh)
    Although modern electronically scanned antennas are capable of generating almost arbitrary near field distributions, antenna designers commonly use analytical distributions that apply only to apertures comprising circles or rectangles. The aim of this Internship project is to undertake a fundamental mathematical study with the objective of finding analytical low sidelobe distributions with given aperture geometry and dynamic range constraints.
  • 3D compressive imaging and sensor development (Selex / Edinburgh)
    Single pixel imaging provides a low cost alternative to 2D arrays, trading the cost of arrays and data read-out architectures for simpler, but slower data acquisition circuitry. The aim of this Internship project is to develop an image recovery algorithm using sparse data sets derived from SPCs or single point scanning systems that could help to increase frame rates. Similar issues arise also in recovering high resolution images with coded aperture masks.
Showing 8 results.
Items per Page 20
of 1