After the authors of the papers had given their oral presentations Mr. ZOUTENDIJK suggested three major points of discussion, in addition to specific questions :
The required conciliation between information needs for financial and operating management and the ways the accounting system has to be changed as a consequence.
The development of heuristic decision: making techniques for daily scheduling purposes, supplementing the existing mathematical programming techniques; the necessity to work towards a combination of human judgement and imagination with the computer's algorithmic power in the form of a manmachine dialogue.
The problem of how to organize and control data processing and operations research.
Then the President gave some reports on the perils of operations research :
to be wrong in defining the end of operations;
to constrain the model of the operations in some unacceptable scheme, which has no other merit than to be well known to the research workers;
to forget, in the conclusions, the conventionality of the scientific language;
to start in O.R. activities with quite complicated problems.
To remember that the modern science began when the scientists gave up trying to resolve the problem of all the Universe and, with Galileo, confined themselves to study the motion of bodies on an inclined plane.
The discussion then centered around the following points: Mr. HUGHES insisted on the fact that in practice management should make use of heuristic methods rather than of a program similar to the one he described.
He also indicated that the present simple model already took several hours of time on a large computer.
Nevertheless, the model could be extended for instance to take account of competitors with different objectives or to include a time element in the reaction of the opponent.
The Petrov presentation was particularly discussed from the angle of the various criteria to be used and of the possible conflict between partial or local optimization and overall optimization.
Mr. STENGEL defined the limits of the programs he described; he pointed out that they are cost minimization and not profit maximization programs, so that they take into account differences in refining costs, but not the influence of different situations on various markets.
The forecasting method described by Mr. MAYER can take account of special events. Inter-refinery shipment of intermediate products and alternative supply of blending components were included in the models.
Answering various questions about the program described in his presentation, Mr. NOGUCHI confirmed that this program is used in actual operation both for long term and for month-to-month scheduling, indicating that for crude costs, he can call on the various sources of the
