I have interspersed some retrospective comments in italics. RJ March 2001.

Editor's Introduction

During the banquet speech at the 1969 Venice conference the retiring President of IFORS, Alec M Lee, spoke of the need for an optimal form of conference and sought support for an experimental structure. He promised that in 1972 the organisers would "stage a different form of conference - one centred on working groups and discussion of major problem areas. Few papers will be prepared and presented and the volume of 'Proceedings' will be primarily a record of the conclusions of the working groups" in the hope that this would prove a better medium for the dissemination of'information among delegates.

The programme committee saw in these remarks a challenge to its powers of innovation - a challenge they readily accepted. Indeed they went further and sought to enhance the value of the conference to delegates by ensuring that papers dealt mainly with solving actual problems. As the chairman, Heiner Mueller-Merbach, put it, this was "not due to any disregard for theoretical work but rather occasioned by the great disparity between the enormous number of really beautiful theoretical developments reported compared to the paucity of original and exemplary applications, notwithstanding the very many problems of public interest currently awaiting the insights of Operations Research".

Given the pragmatic orientation and experimental nature of the conference some brief comments on the novel aspects of the structure are in order.

The conference was characterised by four simultaneous streams. The mainstream was made up of three highlight papers, the "statement of the problem" by the chairman of each of the eight workshop groups, two papers read by representatives of Allied ** organisations, eight state-of-the-art sessions consisting of a central paper and two discussants, and finally the eight reports on the workshop groups. A second stream consisted of the twenty-eight national contributions. A third stream was comprised of the meeting of the individual workshops while the fourth stream was devoted to the discussion forum.

The highlight papers were designed to alert OR practitioners to some of the challenging problems of international and national concern to which Operations Research could be usefully applied. The three speakers from the UK, the US and the USSR presented a remarkably unified approach, given their varied backgrounds, and spoke of the problems of sub-optimisation in national planning. In particular they dwelt on the conflicts between growth and environmental preservation and the needs for interdisciplinary cooperation in achieving a solution. The Russian paper is, in addition, an extremely valuable and vivid statement of the position of OR in that country and the first contribution of the USSR to an IFORS conference.

As already indicated the workshop groups played an important role in the conference. Their objective was to provide an opportunity for delegates to get together in discussion and to think through one of eight management problems of current interest selected on the grounds that they had not been adequately structured as yet. The role of chairman was to guide in formulating the problem by making an introductory statement at a mainstream session early in the conference, to act as moderator during the discussion and to lead the group through an orderly investigation aimed at suggesting a solution, or an approach to a solution. In particular he was required to obtain an informed consensus first approximation solution to be reported back at a mainstream session on the last day of the conference.

Each workshop group was provided with a designated work area and left to organise their time as they wished. Management decision makers, familiar with the problem area in the Irish context or elsewhere, were invited to participate. Conference delegates could join in in various ways: either as full-time "committed" members of a particular workshop, part-time in a "specialised consultative capacity," or on an occasional "quasi-dilettante" basis. Arrangements were made to publish each day the following day's work schedule with details of any special help a workshop might require in the form of "expert opinion" on particular aspects of their problem. Some groups found it helpful to form subgroups to make a more intensive examination of special aspects of the problem under consideration.

Feedback from the conference demonstrates the great popularity of these workshops, and many delegates were enthusiastically in favour of developing and extending the method for future conferences. The active participation of delegates in the discussion lent these activities a greater value than the more passive reception of state-of-the-art papers.

It was left to the workshop to decide what form the report should take. Accordingly a wide variety of formats was employed. No attempt was made to impose any uniformity on them subsequently which may account for some slight unevenness in their presentation.

It had been hoped to publish all eight reports. However, the groups studying supranational problems (*) found their discussions too controversial to provide a synthesis. A detailed report was not supplied by the workshop on education. The delegates attending this popular workshop were convinced of the great merits of the experiment for those who took part, but were less convinced of the value of attempting to condense this experience into a report for the conference. Accordingly no report was presented. The chairman did, however, provide an unusual feedback by asking one of the guests at the workshop to report on her experience. This, together with the chairman's comments on the experiment, will serve to provide some general impression of the topics treated by the participants.

Since the Venice meeting the cause of multidimensional cooperation on an interdisciplinary and international scale was furthered by the creation of FIACC - the Five International Associations Coordinating Committee (**) - in 1970, following a proposal by Professor Victor Broida of Paris - Past President of IFAC. These five associations have agreed to cooperate in their planning and in carrying out certain activities. As a practical demonstration of this solidarity Dr. Zemanek, President of IFIP, and Dr. Cuénod, Treasurer of IFAC accepted invitations to speak on developments in their fields.

The eight state-of-the-art papers are a regular feature of the conferences and need no introduction. The role of discussants was to complement the presentation of the main speaker where necessary. Only seven are reproduced here since the text of the eighth was unavailable from the speaker concerned.

The member societies of IFORS were invited to select a paper or papers for the "National Contributions" session, on the grounds that they were in a better position than the programme committee to review the work going on in their own country and to select the papers likely to prove of greatest international interest. Large societies were invited to contribute two papers (France, Germany, India and Japan) or three in the case of UK and USA. The procedure for selection was left to each national society though a preference was indicated for selection on a competitive basis. In furtherance of the general aims of the conference the national societies were asked to submit, as far as possible, recent unpublished specific applications which displayed a good balance between theory and practice. "Blanket" review papers were not favoured. The quality of the papers published in this section is a clear demonstration of great cooperation between the member societies and the programming committee and an ample vindication of the approach.

It was not possible for the US Societies to follow this procedure by reason of their size and dispersed nature. Accordingly, the three papers from the USA have been grouped into a separate section in the published proceedings as they differ considerably in content from the other national contributions.

In recognition of the fact that there is more to a conference than a series of formal papers the "discussion forum" was devised so as to facilitate the delegates in developing a web of informal direct contact with people sharing their particular interests. The "forum" consisted of a number of booths equipped to accommodate small groups. In addition two lecture rooms were available for those topics for which a large attendance was anticipated. In keeping with the purposes of the experiment only a proportion of forum time was scheduled in advance of the conference. The remainder could be reserved by giving the forum manager notice by lunch time of the previous day. This was to ensure their receiving adequate advance publicity in the forum bulletin for the day concerned. These proceedings contain a report on this experiment, given by the manager, RHW Johnston.

Acknowledgements

Under the chairmanship of H. Mueller-Merbach of Germany the programme committee comprised: RHW Johnston, Deputy Chairman - Ireland, E Johnsen - Denmark, BM Brough - UK, G Morgenthaler - USA, FG Foster - Ireland, A Straus - Switzerland.

The eighth member, Paul Naor of Israel, died in an air crash while travelling to a meeting of the programme committee. In acknowledgement of his work for the conference meeting number 10 was dedicated to his memory.

Under the chairmanship of FJ Ridgway the Irish committee responsible for local arrangements comprised:

DO Byrne (Secretary), T Forsyth (Treasurer), M Flinter, MA Foley, DM Kennedy, JM Lynch, J McGilligan, S O'Carroll, WP O'Grady and M Ross.

In the second half of the period this committee benefited greatly from the experience and dedication of Miss T Higgins who became full time conference organiser and head of the Conference Secretariat.

Both committees profited considerably from the experience of the two IFORS officers who were ex officio members of the programme committee: A Jensen, Denmark - President, AM Lee, UK - Immediate Past President, while the advice of the IFORS Secretary, Mrs M Kinnaird, and the Treasurer, RH Colcutt, was invaluable.

For my part, as editor of the proceedings, I should like to record my thanks to John Lawrence and Einar Frederiksson for their helpful advice. Because of space constraints I found it necessary to ask some contributors to shorten their contributions. I would like-to express my gratitude to the many authors who responded so readily to my request and who were so prompt and courteous in their dealings with me. Finally, I would like to thank the Director and Council of the Economic and Social Research Institute in Dublin who graciously released me from other duties to enable me to devote more time to the production of these proceedings.

Fé rath Dé go rabh siad uilig.

Notes

** i.e. IFORS, AICA (Association Internationale du Calcul Analogique), IFAC, (International Federation of Automatic Control), IMEKO (International Measurement Confederation) and IFIP (International Federation for Information Processing).

Simulation and Validation: Comment by RJ on Naylor Paper

(1) The ratio of simulation to analysis in a model can have any value between zero and one; it is not a binary variable.

(2) An experimental structure, which is a model of a real system, ought to be designed so as to produce a big signal, well above the noise level, if insights are to be gained.

(3) There is a need for some objective common measure in multiple-response situations; subjective measures, such as utility, are of doubtful value.

Dr Naylor gives the impression that he regards simulation and analysis as two quite separate camps. I suggest that it is possible to build hybrid models, containing elements described analytically, interacting in response to signals one or more of which are simulated.

If I may take an example from my own experience, in the early days of real-time airline reservations systems, (1963-64), little or nothing was understood about the statistics of the response of random access computer systems to an environment which manifested itself as a stochastic demand for service of randomly varying type. Serious errors were made in the design of real-time systems as a result. Simulation experiments were done subsequently to try and discover the nature of the design faults, and valuable experience was gained.

Round about this time I was working for an airline and I produced an analytical simulation of a real-time system, in which I used the results of queue-theory to predict waiting-times consequent on the demand for services for the central processor, for the channel serving the files and for the files themselves. The demand for services for each of these elements was calculated from a knowledge of the number and type of "messages" which were "in the system". "Messages" were specified by patterns of demand for service by the central processor and files. The "next message" was chosen by the one "Monte Carlo" procedure in the model. This "analytical simulation" gave results in broad agreement with the traditional simulation developed by the manufacturers; it predicted the same bottleneck, namely, the file-access channel. Multiple channels were subsequently introduced. The manufacturers' simulation took some man-years to develop, and hours to run. The analytical simulation took some man-months, and minutes. (This work is on record in the 1965 AGIFORS proceedings at Chicago.) I was convinced by this experience of the wisdom of practicing the maximum economy in the use of random variables.

I turn now to my second point: the need to design the experiment so as to produce a big signal. Nature has provided an abundance of noises with which the signals describing the states of our systems are masked. The philosophy of experimentation which I imbibed when working as a physicist consisted in constructing an experimental model of the system which reduced to the minimum the extraneous noise and focused on those signals which were considered to be the key to the essential dynamics. I get the impression that the type of simulation philosophy expounded by Dr Naylor is at variance with this: he feels it necessary to make the noise-level of the model system comparable to that in the real system, so that he is in the happy position of being able to use all the tricks in the statistician's bag to pick the signal out. When 1 was working in physics, the lore used to be that if a man had to resort to sophisticated statistical techniques, his experiment was suspect. This may be seen by some as an extreme view, but it is useful to bear it in mind when setting the experiment up.

Fortunately, in business systems nature has provided a filter with which the experimenter can pick out the relatively small number of significant variables from the mass of background. This filter is the judgement and experience of the manager: those who have not developed a feel for the significant variables from the experience of working the system are unlikely to have survived. Thus the "problem of many factors", while theoretically present, can in practice often be avoided.

It therefore becomes possible to make meaningful simulations of complex systems which give large and clear signals if we allow ourselves to use judgement and experience in the suppression of irrelevant noise. I, personally, prefer this road to that mapped out by Dr. Naylor. If this restricts the class of problem open to me, I accept this restriction. I have no doubt that there are many who would reject it. I suggest that the ultimate arbiter is the client, who will pay for that which gives him most insight per dollar.

Finally, I feel I must contribute to the discussion on multiple responses and Dr. Naylor's proposed approach via utilities. I would like to see someone working on an alternative approach in which the sole measure is the survival probability of the system as a viable organism. There are insights to be gained from current work on ageing in biology; this depends on information-theoretic concepts such as garbling of coded data, entropy levels, etc.

I suggest that an approach to the problem of viability of an economic organism can be made via a thermodynamic model, with temperature and entropy defined in information-theoretic terms. One can envisage an economic organism (a firm) ingesting nutrient from a disordered environment, ordering the ingested raw material into finished products, which are placed with precision in a disordered market. Each step involves entropy reduction, which costs money. It is useful to reflect that the product of entropy and temperature has the dimensions of energy, which is the same thing as money. Each step in the process has an entropy change associated with it; this entropy change occurs at a "temperature", which may be conceived as a measure of management ability. A "hot" management can reduce entropy rapidly.

An economic organism is viable if its revenue from sales exceeds its total costs; the cost function includes "volume * unit cost" terms and "temperature * entropy" terms. This approach, therefore, contains the embryo of a theory of management costs or "overheads", as the core of its approach to the question of viability of the economic organism. I commend this approach to the theoreticians. It is, as I said, related to a body of theory which is developing in biology with sound roots in physics, thermodynamics and information theory. It represents a radically different approach to that of utility theory, although both.relate to measures of value. It substitutes for a multiplicity of subjective sub-goals a single over-riding goal, survival; this goal is of very deep significance and has deep roots in biological and sociological evolution.

To my eternal regret I never managed to develop this promising theoretical approach to the stage where it would provide useful quantitative tools, though it has been a source of qualitative insights. RJ 2001.

Finally, may I make a plea that the theoreticians should pay more attention to the challenges thrown by by the practice of OR. The theory-practice ratio in OR is out of proportion to that which obtains in other branches of applied science. Because theoreticians are solving problems posed by each other and neglecting the experimenters' world, important areas such as that outlined by Dr. Naylor are neglected. Consequently their development is stultified and pushed into repetitious applications of statistical filtering of dirty signals, which we have all seen before. We need more insights into the inside of the black box, illuminated by good theory. We need.more analysis in our simulations. We need to design our experiments so as to keep the noise level down, and finally we need to develop an integrated theoretical approach to the measurement of system viability which could possibly be based on thermodynamic concepts.

Report on the 'Discussion Forum'

1. Introductory Note

I cannot guarantee that the attendance figures given are accurate or that they reflect any meaningful measure. It is, however, interesting to analyse them. The attendance figure is given in brackets after the name of the contributor. In some cases an alpha character is substituted. The meaning of these is as follows:

x Author was registered at conference but was unable to be present;

y Location of meeting was changed from the forum to a lecture-room elsewhere, leaving us with no attendance record. We can surmise that it was large;

z We have no record of the attendance due to negligence of the duty officer.

Some authors submitted abstracts in advance, were scheduled, but then did not turn up. In one case, that of the late Dr J Walsh, it must be recorded with regret and some shock that he died during the course of the conference. In other, less forgivable, cases the scheduled authors were not even registered at the conference; one was present but refused to register. Where people turned up for an absent author, we have passed on (out of sheer goodwill; we were under no obligation to do so) the names and addresses of enquirers to whatever contact-point we had.

We would like to place on record for future conference organisers that we think this practice reprehensible, as it makes the avoidance of schedule conflict more difficult by increasing the event-density, and generally increases the "noise-level" in the conference management system. Thus, if a prospective participant is uncertain as to his coming, he should make use of the real-time scheduling facility, to allow for which the forum was designed.

Finally, the grouping presented here is my own; it seemed worthwhile to look at the phenomena of OR fashionability. I am impressed by the high level of interest in models involving judgement and the human element; I am not surprised at the relatively poor support for applications of traditional techniques; I am pleased at the suggestion of a keen interest in social and environmental problems (although the numerical data in this case seems to say otherwise, there were two "y" papers, and the "zero" was a late submission, so I am inclined to give it little weight). I am surprised and disappointed at the poor support for the "development areas" group: most of this was clearly an effort to build bridges into disciplines where the relevance of OR is hitherto unrecognised. There is scope for more interaction with the civil and electrical engineers, and other such disciplines with long and distinguished numerate traditions; many OR techniques in fact have independent and parallel histories without the OR label.

On the whole the above comments have stood the test of time. OR as an active branch of academic applied maths, with emphasis on theoretical refinement, has survived in 'ivory tower' mode, to the extent that people doing what I always regarded as 'real OR', ie addressing and structuring real-world problems and coming up with (perhaps fuzzy) quantitative 'what-if' analyses as an aid to their solution, taking into account the human dimension, have gone off and done this without the OR label.

2. Application of traditional OR techniques

2.1. R.L. Boase (3); 6238 Mercedes, Dallas, Texas, USA

A method for validation of multivariate simulators whereby statistical tests are applied to a pair of data sets, (one real, one simulated) which need not be continuous, random or independent but must possess joint null distribution symmetry.

2.2. Major K.E. Brant, USAF (1); Aeronautical Systems Division, (ASDISDMC) Wright-Patterson AFB, Ohio 45433, USA

Military applications of Bayesian statistics.

2.3. T.A. Ciriani (x),. Pisa Scientific Centre, Pisa, Italy

Post-optimal analysis of linear-programming problems by free parametric variation of the coefficients.

2.4. Kurt Haessig (2); Inst. fur Operations Research, Univ. Zurich, Zurich, Switzerland

A system of linear equalities or inequalities, including linear programmes, is applied to problems in banking and in multi-commodity flow.

2.5. Pierre Hansen; Inst. d'Economie Scientific et de Gestions, Lille, France, and Dirk van Oudheusen; Univ. of Brussels, Brussels, Belgium. (10)

An implicit enumeration algorithm is proposed for batch processing of jobs on a production line where the latter can exist in several states with different costs, and where sequence-dependent costs associated with changes in state can also be incurred.

2.6. CP. Saksena (6); Computing Lab., Univ. of St. Andrews, St. Andrews, Scotland, UK

A modification of the bounding hyperplane method of linear programming is developed which uses the equality constraints to obtain an efficient optimum.

2.7. Shafik L Saleeb (6); Inst. of Statistical Studies, Cairo Univ., Ghiza, Egypt, UAR

A simulation technique, enabling distributions of CPU time, I/O events, core requirements etc. to be determined for a computer, has been used to improve the operation, and plan the expansion, of the Cairo University system.

3. Internal OR politics

3.1. J. Abrams (6); Inst for the history and philosophy of Science and Technology, Univ. of Toronto, Toronto 2B, Ontario, Canada

The implementation of operations research in Canada.

3.2. P. V. Norden (12); Data Processing Div., Columbia Univ., White Plains, New York 10604, USA

It is suggested that the communication gap between management science and top management may be bridged by an Internship Programme sponsored by TIMS.

The relative popularity of the above Norden paper, and indeed the papers in the next section, indicates a pioneering awareness of the importance of human and organisational issues in management science. RJ 2001.

4. Planning models involving judgement and the human element

4. 1. A. Buttimer (22); Princeton 1, Ardilea, Roebuck, Dublin 14, Ireland

A low-storage computational system utilising both empirical and exponential smoothing terms, suitable for short-term planning and forecasting, has been developed which in the application proved 35% better than best known alternative.

4.2. H. Muller-Malek and J. de Cock (14); Fac. of Economic Sciences, Univ. of Ghent, 9000 Ghent, Belgium

The behaviour of a known decision-maker with regard to an operational repetitive task in an industrial environment was simulated on a computer; results were compared with the output from a mathematical programming model of the same problem. Satisficing behaviour, it is suggested, should not be considered as a deficiency in human behaviour, rather as an efficient way to deal with compromises.

4.3. H.J. Rasmusen (13); Inst. of Mathematical Statistics and Operations Research, Tech. Univ. of Denmark, 2880 Lyngby, Denmark

A corporate planning model is proposed in which each decision-maker within the organisation has his own sub-model; it is suggested that by this means conflicting objectives might be exposed, and internal communications improved.

4.4. A.M. Sutton (16); Atkins Planning, Woodeote Grove, Epsom, Surrey, UK

The use of modelling systems in budgeting, variance analysis and capital investment analysis.

5. Social and environmental problems

5. IS Basheer Ahmed (4); Univ. of Western Kentucky Bowling Green, Kentucky, USA

A common numerical measure is suggested for environmental effects on water, air and land, relating the quality of human life.

5.2. Werner Burckhardt (2); Systemberatung Fides Dusseldorf, Dusseldorf, Germany

A model for planning teacher supply is proposed which allows for demand, university capacity, retirement and sabbatical years; the discrete maximum principle of Pontryagin is used.

5.3. ND Francis (0); Dept of Computer Science, Trinity College, Dublin, Ireland

An educational system is modelled by a linear matrix differential equation; control-theoretic concepts enable a dynamic programming approach to be developed, whereby inputs of student, staff and space are optimised to achieve a long-term educational target.

The poor attendance at these 'manpower planning model' presentation perhaps is an indicator of the trend away from 'optimising' and towards 'satisficing' and paying attention to the importance of the human element. RJ 2001.

5.4. Sir G. Goodeve; Tavistock Inst. of Human Relations, London, England; and C.J. Hitch, Univ. of California, Berkeley, California, USA (12)

An informal discussion took place arising out of two of the "Highlights" papers.

6. Development Areas

6. IM Citina (0); Inst. Matematico Polytechnico, 10129 Turin, Italy

A dynamic programming technique is used to generate a pressure law which minimises reaction time in an industrial chemical process taking place in a discontinuous reactor.

6.2. JH Langston (4); Social Systems Program, GEOMET (Inc. Rockville, Maryland 20850, USA

A neighbourhood health centre is studied; the effective use by target population, the degree of satisfaction and the service-level are determined; organisational and functional characteristics of the centre are related to its effective performance.

6.3. Frank H Trinkl (8); Graduate School of Public Policy, Univ. of California, Berkeley, California 94720, USA

Resource allocation in social programmes is modelled in a manner permitting consistent decisions within a hierarchy and using consensus methods when judgements concerning value and effectiveness differ.

6.4. Christian Dyrberg (1); Inst. of Mathematical Statistics and Operational Research, Tech. Univ. of Denmark, 2800 Lyngby, Denmark

Renewal theory and other stochastic processes were used to determine risk profiles for various alternative methods of constructing a harbour.

6.5. TN Manias (6); Research Inst. of Ancient Knowledge, c/o D. Xirokostas, 24 Armatolon and Klefton Street, - Athens T. T. 704, Greece

It is suggested that the proportions of the pyramid of Cheops may be related to some mathematical properties of sums of series of prime numbers, and that its absolute scale may be determined by reference to certain distances between centres of importance in Greek antiquity; a prior Greek influence is therefore suggested.

This was the paper which attracted press attention; it was somewhat egregious, and indeed verging on the eccentric, a hazard to which the 'poster sessions' of conferences are open, with the peer-review effort being concentrated on the mainstream papers. RJ 2001.

6.6. Paul R Milch (1), Naval Post Graduate School, Monterey, California 93940, USA

A need is suggested for development of queue theory to cater for cases where a finite number of customers (say less than 20) form a queue for a relatively short time: in this case the queue length distribution is of no interest but the total system lifetime and the mean customer-in-system time are important.

6.7. W Riley (4); MISIDP Unit, ILO, Geneva 221 Switzerland

Survey statistics and personal experiences from selected developing countries in Asia, Eastern Europe and Africa indicate a near crisis situation in respect to the application of computer technology contributing to national development. The alarming rate of increase in computer hardware acquisition, doubling the number of installations every two years, matched with inadequate or non-existent training for management and computer specialists is resulting in underutilization of the available computer capacity on non-significant user-applications. The background information supporting these conclusions are presented, in order to promote the exchange of experiences by others, leading to a discussion of possible steps to alleviate the obstacles faced by developing countries in utilizing computer technology for their national development.

The Riley paper was an indicator of the acuteness of the 'core-fringe' problem which is still with us, and which extends across all domains of science and technology. The above abstract was badly worded; he meant that the capacity is underutilised, and the applications where the capacity was used were non-significant.

Some navigational notes:

Copyright Dr Roy Johnston 1999