ASSUME 2000 Meeting Report

The meeting took place at Coventry University on 11th April 2000 with 32 participants drawn from both academia and industry. Despite the miserable weather outside, the day turned out to be both informative and enjoyable.

The first speaker was Barrie Baker of Coventry University who talked about a variety of spreadsheet applications in introductory operational research.  He went on to introduce the audience to the Auditing Toolbar and demonstrated a number of ingenious applications of this feature, including critical path analysis and depot location.  The spreadsheets are freely available from Barrie's research pages on the Internet.

Tony Greenfield, an independent consultant, then spoke about a short course in design of experiments that he has developed for production and development engineers.  The coursework assignment for this course has been set up in such a way that each student is allocated a random set of data for analysis.  Tony very kindly handed out a copy of the coursework, together with a disk containing the Excel spreadsheets, to everyone at the meeting.  The various files have been zipped and can be downloaded  from the Internet by right-clicking here and choosing the appropriate browser option.

The morning session was rounded off by a demonstration of Unistat 5 by Mehmet Toker of Unistat Ltd.  The Unistat package can be run as an Excel add-in, but uses its own algorithms to guarantee accuracy. A number of impressive new features were displayed.  A free demonstration version is available from the Unistat homepage.

There was some discussion over lunch about the ongoing problem of the bugs within Excel's statistical facilities.  The ASSUME web-site still offers no guidance on this.  It was felt that we should make an effort to complete such a resource.  A helpful contribution was made by Alessandro Leidi from the Statistical Services Unit at Reading University who distributed three helpful advisory booklets on the use of Excel for data management and analysis.

After lunch Geoff Morgan from Nag Ltd spoke about the second generation of Nag statistical add-ins for Excel. These add-ins are notable in that the output generated is dynamically linked to the data.  There have been a number of significant improvements to the user interface, but work is ongoing to make it even more user-friendly.

Francois Sermier from CEPE in Paris spoke about using Excel for health insurance forecasting.  He described a training course that he is involved in for the Caisse National d'Assurance Maladie, whose regional statisticians are required to make regular forecasts on literally scores of series.   The Holt-Winters method is used extensively and Francois showed how this is implemented in Excel, and there was some discussion regarding problems such as structural change.

The day was brought to a conclusion by Jim Shalliker from Plymouth University who spoke about Excel's Solver.  Jim described some of the alternatives to Solver (see Frontsys and Lindo) and presented a series of examples of applications of Solver for linear optimisation.  He concluded by giving a brief Use of uninitialized value in concatenation (.) or string at E:\listplex\SYSTEM\SCRIPTS\filearea.cgi line 455, line 73. outline of each of the Solver options, with advice on suitable settings depending on circumstances.