Discussion of the Work Packages

Location: Cascais, Portugal
Date: May 8-10, 2002

Notes to Work Package Leaders

The following are generic deliverables and milestones relevant to work-packages 1,2,3, 4 and 7

D.1 at least 2 General workshops and seminars that will provide opportunities for inter-disciplinary information exchange
D.2 at least 2 Industrial Visits aimed at fostering new research collaborations
D.3 at least 2 Industrial Visits aimed at demonstrating best practice
D.4 at least 2 case study based journal papers
D.5 at least 2 articles in the European Popular press

M.1 2 Industrial Visits completed month 18
M.2 2 Journal Case Studies submitted month 24
M.3 2 Workshops completed month 30
M.4 4 Industrial Visits Completed month 30
M.5 Paper articles published month 36

WP1 - Design of Experiments

Work Package Leader - University of Warwick

Dedicated to furthering the use of DOE in an industrial and commercial context. The emphasis will be on developing concepts, to be of use daily in quality and productivity improvement.

Description of work/tasks Organising workshops, meetings and events. Publishing papers, articles, notes. Organising network meetings on topic.

With these objectives in mind the following was discussed.

A Committee has been set up which includes:
Henry Wynn
Rob Deardon
Eva Riccomagno
Allessandra Giovagnoli
Oyvind Langsrud
Stefano Barone
Dimitar Vandev
Ulrike Groemping

- Matforsk (month 22) workshop would be oriented towards the food industry
- DEINDE (proposal to co-join with a workshop in Cagliari) Sardinia May/June 2003, D Romano will be helping with the organisation. Levi Raffaello will advise and notification will be posted onto the web
- Suggestion to broaden the participation to members of the Pro-Enbis group, ENBIS membership and to industry. To open up attendance to the service sector. DEINDE have a major impact on experimental design in Italy. Anticipated there will be 40 -50 participants, approx. 1/3 are from industry. Further suggestion to expand the duration of the whole event from 2 to 3 days.
- Cagliari event could be service workshop oriented towards experimental design/DOE in call centres and the tourist industry
- Noted also that different level workshops could be run for example "road-show" type event to be held at a local hotel; these events would be free of charge although it was pointed out this could lead to a lack of commitment from the attendees.
- It was noted that it is important for the workshops to have a clear identity aimed at particular sector issues.
- Warwick date TBC
- Other activities include a poster session for case studies in Rimini.
- Case studies will be placed onto the web site
- Preliminary case study protocol - Committee to prepare a consultancy document. Submit case studies following this protocol. Key problems are to be practical, translating data to actual work in the company - Agreed this is a crucial area and would be further discussed in Rimini
- Immediate work - need to get case studies. Need to discuss workshops/seminar plans and a definite
- E- mail contact list. Pulling in existing research. Pull in existing papers for the DOE work-package.

Industrial Visits
Levi has received a contact from a MD of a company to develop experimental design in the company. 2 differing styles are a possibility, firstly an individual visit to the company and secondly a road show. Agreed to develop a pre-visit pack. Objectives of the industrial visits - conversion to techniques and to set up a long-term relationship with the statistical community.

WP2 - Data Mining/Warehousing

Work Package Leader - Eurandom

Dedicated to exposing researchers and industrialists in the field of statistics to data mining problems and tools, so that the natural skills of statisticians, such as the ability to model real observational data, can be applied to data mining as well.

Description of work/tasks
Organising workshops, meetings and events. Publishing papers articles notes. Organising network meetings on topic.

With these objectives in mind the following was discussed:
- Wim Sendon has agreed to take over project management. Not much happened yet but will.

Eurandom will jointly fund a major workshop, it was noted that Eurandom can pay for things that the Pro-ENBIS project cannot for example to pay for attracting "big names" from elsewhere if we need them. (Friedman etc) they may also have monies for post doctoral involvement to facilitate data mining at a research level

Issue of data on integrated networks - to have links with major important data sets e.g. genomes could be linked/ deposited, data depository, house the data locally in Eindehoven. There is the important issue of the housing of the sheer magnitude of large data sets

There will be data mining parallel sessions at the Rimini conference

Data mining committee will include:
Henry Wynn
Wim Senden
Alessandro di Bucchianico
David Rios Insua
Andrea Alhemeyer-Stubbe
Paolo Guidici
Carla Brambilla and
Irena Ograjensek (agreed)

Data mining techniques - different subjects under data mining include such areas as decision regression trees, neural networks, genetic algorithms and dynamic data mining; with data coming on line fast in real time with a large a magnitude of difficulty; this is definitely the future, The companies that address this will be the most competitive in the future.

Andrea has volunteered her extensive data mining networks.

Further discussion took place with breakout group including Andrea, Ron, Irena, Henry, Rob, Jorge and Bruce. Discussion included
- Additions to Committee Membership
- Workshop at Eurandom
- Literature, Papers, depository, industrial visits
Contacts with other societies

Industrial Conference in Data Mining to be held in Leipzig on 13th - 15th June 2002. The conference is aimed at people carrying out data mining in industry. Would be of direct relevance to Pro-Enbis and ENBIS working group in Data Mining. It was agreed to use the conference for the DM group to meet again. Further information can be obtained from http://www.data-mining-forum.de.

Potential invitees include Friedman, David Hahn and Wise
Other workshop activities: Eurandom workshop to be held in June 2003 (format will be English)
To meet again in Rimini in September, Paolo Guidici would be contacted with regards to Data mining activities

Presentations in Data Mining to Industry
- Template for presentations in DM to industry to be set up
- List of DM websites
- Management level of DM concepts (grab their interest Euros)
- To concentrate on companies that outsource DM (which software/invite vendors)
- Vendor day in Eurandom suggested
- Rimini to invite SPSS/SAS/Oracle to stress to software companies the particular interest in DM
- To hold a session on Customer Relations Management

- Issue - how to house huge data sets
- Organising so 'ordinary' people can 'play' with data sets
- Perception of ease of result, spend Euros (complexity)
- Graphical models - interfaces
- What they as customers get out of it, to discuss data with customer and the success of DM
- Catalogues - cost issue (customer could make massive savings)
- Encourage DM with existing software - DM with their PC's
- Automatic Model Generation - alternative definition of DM
- Audience 2 distinct markets and how to address them
- Look for deposits of decent data so we can use (excluding marketing)
- Get process data (e.g. time series, multi correlation)

Industrial Visits
- What do we do?
- Describe technology + enlighten
- Two-way learning (what we can learn from them for example Nestle case study)
- Demonstrating Best Practice
- Fostering research collaboration
- To wear Pro-Enbis hat but also we are experts
- Industrial visits to be carried out by Andrea and Jorge
- ENBIS Data Mining Road Show - stress we are NOT selling anything; Free of Charge; Pre-consulting
- To aim at people who are not processing data at the moment e.g. banks, airlines, insurance. To think of possible companies. Suggestion to contact ENBIS corporate members in the first instance
- Data collection on web transactions - e business (electronic sales model)
- To invite John Brocklebank - SAS guru
- To approach Mercury, Shell Global Solutions and utilities

Work Package 3 - General Statistical Modelling, process modelling and control

Work Package Leader Greenfield Consulting

To improve the use of process monitoring and control methods for both industry and commerce and help companies introduce advanced techniques, via workshops and visits.

Description of work/tasks
Organising workshops, meetings, and events. Publishing papers articles notes. Organising network meeting on the topic

The work package group had met the previous day and discussed ideas and reviewed progress. Here are the main findings:

- Workshop development including a base of what to say when carrying out works visit, developing papers and reports to the trade press.

- The workshop would address all levels and range from a 1/2 day introduction to full 2 days.

- Industrial process model - simulation, application and context.

- Process Simulation Model - had been developed and would act as a doorway to statistical methods in business and industry, Xavier Tort Martorell, Tony Greenfield and John Shade were authors and other people are contributing. It was envisaged that the model would help in solving business problems, be practical and of use in industry. The model would assist in telling the clients how simulation models can help in areas such as Statistically designed experiments, aims for best setting and processes, identifying problems and help give solutions. It would include examples of manufacturing problems for example breakdowns, poorly trained staff, difference in inspector's etc...
The graphic example given was a game about the production of aluminium wheels (suggested by Xavier - Luis Marco wrote the game)
Context - we need to be able to go into industries with a simulation and to tailor the model to different contexts e.g. chemical, mechanical, electronic industries, assembly and test and involve case studies in the relevant field. It was further suggested that we have some games that avoid the context too.

- Web Library - which would support the whole of Pro-Enbis activity and would include:
Case studies
DOE games
SPC games
Production Planning

- From the DOE /SPC game we could produce tutorial material and software options, to include also data mining and reliability demonstrations and teaching multivariate control.

- Commercial application - CD-ROM will be freely available for pro Enbis and Enbis members there are no restrictions to use but are not to be sold or edited. To acknowledge author name Encourage during industrial visits and training courses

- Further contributions were invited from all; it was agreed the simulations would be extremely useful. To send case studies, games and simulations to Tony Greenfield

- Notes from further discussion noted that we could possibly stream games live onto the web to make it more dynamic

Work Package 4 - Reliability, Safety and Quality Improvement

Work package Leader - IST, IMPRO

To disseminate help on improving the reliability of processes, products and systems and help facilitate the philosophy of statistical thinking as applied to the world of work via workshops and visits.

Description of work/tasks
Organising workshops, meetings, and events. Publishing papers articles notes. Organising network meeting on the topic

Discussion centred on the following:

- There will be a workshop held in conjunction with ESREAL 2003 (European Safety and Reliability Association). The conference includes tutorials and workshops both advanced and basic and round table discussions, Next year's conference could involve Tony's process simulation models
- A workshop to be held in Slovenia to include a 'clinic' for companies was suggested
- Industrial visit to Trondheim, focusing on reliability has been arranged.

Case Based Reasoning Library to be built upon
- Stochastics for the quality movement to cut across all work packages. To consider what are the main issues, what kind of industry do we help? Library to match the problem with the solution, to establish what is the state of knowledge at the moment in our industry. To identify sectors to ascertain their needs, there will be areas that we could fulfil immediately but others would need to be bettered. What does industry need in particular circumstances and in particular problems?
The following sectors were identified in the breakout session:
Food - Oyvind Langsrud
Ship - Oystein Evandt
Semi-conductor - John Shade
Electrical Engineering - Antonio Pievatolo
Water - Dave Stewardson
Telecommunications - Fabrizzio Ruggeri (?)
Ron Kennet to provided case studies for food, chemical, pharmaceutical etc
Pipelines /gas - Shirley Coleman
Gas pipeline - perhaps Shirley and Oystein to work on this together
Shirley Coleman asked for a template of what Maria is looking for
John Shade already has a template - will send to Maria to pass to Shirley
Antonio Pievatolo asked Dave Stewardson what could he promise to organisations to convince them about the benefits of providing this information
Dave Stewardson - if Antonio Pievatolo has some companies who he thinks we can visit we can help them
Jukka Salmikuukka has a few companies who may be interested in an industrial visit
Dave Stewardson suggested organising these visits
Dave Stewardson proposed if anyone has companies who want a visit - we will put together a plan to present at each of the visits
This is how Trondheim occurred
Who else is interested in going to Trondheim 21 June?

Creation a Reliability Programme
Ron Kennet - motivation to get involved does not come from the reliability of the product
Maria asked Ron to send information on Reliability programmes
Jukka Salmikuukka has a case of Life Cycle Process related to Marketing
Maria asked Jukka Salmikuukka to send this to her
Maria then asked for other examples on other areas of reliability

Repairable Systems
What are people publishing in this area
Do we have any suggestions of where most up to date information in this area for Repairable Systems?
Ron Kenett referred Maria to the American Society of Quality at ASQ.org to review reliability section

Shirley Coleman asked whether other subject areas in this work package would be addressed later. It was envisaged that they would all link together, firstly focusing on reliability issues then address how they link with other areas such as the quality movement

Advise Maria of the companies we are involved with and if you are interested in collaboration on this area. KPA ltd (Ron Kenett) noted that he might have information for the case based reasoning library. A reference list would be put onto the web.

Reliability discussion centred on the issue of practitioners in the field applying the findings or statistical methodology in the real world; issue of whether there is a divide between the theorists and the maintenance practitioners. There is a real issue to be addressed around reliability in industry, is it a communication failure or a failure of application, has the message of the importance of reliability been lost. It was noted that some companies do not carry out even the most rudimentary reliability methods (histograms, hazard plots). We need to demonstrate to practitioners the practical, real information they can get from data (make it relevant to them)

Work Package 5- Discovering European Resources and Expertise

Work package Leader - ISRU

This work package was not discussed fully at this meeting. The group's attention was drawn to the first milestone, which is to establish a list of journals by month 12. Oyvind Langsrud advised the group that a survey was currently being carried out in the USA and would comprise of a list of journals (checking this out with Oyvind for clarification).

Work Package 6 - Drafting and Initiating Further Activity

Work Package Leader - ISRU

Attention drawn to the milestone of month 6, to draw up a list of potential partners for European Training Networks. All present were asked to have input into this list; to advise of potential partners, different types of training and delivery. To advise Co-ordinators (Lesley Fairbairn).

Work Package 7 - Management Statistics

Work Package Leader IBIS

Dedicated to furthering the use if statistics in strategic planning, operational control and marketing in an industrial commercial context. The emphasis will be on developing concepts for ease of application within the management capabilities of European Industry

Interested parties:
IBIS UVA (Ronald Does)
University of Ljubljana (Irena Ograjensek)
UNICAN (Alberto Luceno )
Nick Fisher
Soren Bisgaard.

Joint meeting ISI and ENBIS in Barcelona. Thursday & Friday w/c 21st August 2003, we anticipate our own activities before the main conference. Xavier Tort Martorell and Albert Pratt (ISI) will be involved as local organisers. Committee to be agreed.

Soren Bisgaard will be in Amsterdam for 6 weeks shortly it is anticipated that a lot of work will be done.

- It was agreed to hold a workshop in Key Performance Indicators in the future and one aimed at the service sector
- Workshop in measurement of service quality would be organised: following and evaluating trends in business environment (with special emphasis on prognoses of future developments of technology)

Feedback from discussion paper on Management Statistics circulated by Irena and Ronald, question raised was the paper an accurate reflection, had the issues been covered.
Three levels of Management
strategic management
tactical management
operational management

Feedback from 'round robin'
How do you relate to the various levels?
Pg.2 management system - not tampering to not produce extra noise.
It is an important issue that top managers need to get enthusiastic about statistical methods. Couch benefits in terms of money, what are the real benefits,
Six -sigma is worth taking into consideration.
There are 3 aspects but do not get pushed into these pigeonholes,
Evidence based management.
Like to see focus on DOE in non- manufacturing for example sales staff in shops KPI's various weighting changes in relation to the market but are not good for understanding a society and statistics
TQM systems
Need to think about techniques, can't focus on what we see as statistics
Easy to model onto the theory, the modelling becomes easier as the levels get lower.
Also becomes more and more easy to abuse the statistics.
Need to distinguish between management statistics and mathematics, prefers evidence based statistics
Issue in management is to accurately break down performance measurement Relevant publications to be advised
Reasoning is becoming more based on numbers (e.g. employment satisfaction) there is opportunity for statistical applications
Modelling issue what models do you use at different scales - is a critical issue
Focus on SPC
We have to show that this translates into the bottom line
Advantage of data management applications
Government and areas of the economy is a relevant area (universities etc)
Digital dashboard for enterprise, Nick fisher will be here at the Barcelona meeting and could demonstrate his thoughts there
Managers as the customer -need to match how managers manage and what we can offer them.
There are things that can be done in improving the models
The balance score card here is incomplete what you want is a cause and effect relationship the aggregation issue is almost secondary.
Real option models - financial techniques applied to real management, an opportunity to be more relevant.
Decision making focus - Where and why people may be dissatisfied with their decision making.
Bogsat decision making - bunch of boys and girls sitting at a table
Need to encourage a degree of dissatisfaction with the decision making process.
D wheeler managing chaos -
M Harry book on six sigma
Top managers look at very disaggregated data
The combination of knowledge and leadership are exceedingly rare
SME's have all of these categories of manager in 1 person
Decision support systems to conform to regulations
Stakeholders all the relationships the company
Statistics and evidence based management
Need risk related KPI's
Logistics in the supply chain to be considered
To have feedback from the Managers
To illustrate how the principles here apply to the running of Pro-Enbis.

Ronald and Irena would consider the feedback from the discussion paper and will collate the references. The resulting paper will be submitted to a journal

Long term target write a book a Statistical Road Map for Europe; this could be included in the state of the art report required to the European Commission

Schedule of workshops

WP1, WP7 June 2003 Cagliari
WP4 June 14 (evening)
June 15 (all day)
Pre-ESREL (June 15-18)
WP2 May 2003 EURANDOM with PE meeting
WP3   Preparing material first
WP2, WP7 September 25-27 Rimini.
3 days' course on
Statistical methods for evaluating
the Effectiveness of advertising
WP1 May 27, 2002 Varna

Back to meetings.