It may not be possible to be wholly objective, as Messrs. Stimson and Thompson (among others) assert (see [Stimson, D. H., R. P. Thompson. 1975. The importance of ‘Weltanschauung’ in Operations Research: The case of the school busing problem. Management Sci. 21 (10, June) 1123–1132.]). Nevertheless, one should be as objective as possible in OR, or any other science, and this carries with it the concomitant duty of accurate and adequate descriptions—including descriptions of work by others to the extent that it may be pertinent.

The papers which follow have been assembled from two sources: the ones by Dr. Branch and by Professors Eastman and Kortanek were submissions to Management Science; the others formed one part of a program organized by Ezra Glaser for the American Association for the Advancement of Science's 1969 meetings in Boston. As vice president for AAAS' statistics section, Mr. Glaser had organized an all-day session on “Statistics, Governments and the Analysis of Social Problems.”

Presidential address about the background leading to TIMS formation. Delivered before the first national meeting of the Institute of Management Sciences, Oct. 21–22, 1954. Parts of the address are omitted.

Bell System applications of management science are reviewed, with examples, under three categories: (a) Facilities—providing the physical communications plant, (b) Operations—concerning both the plant and the organization, and (c) Marketing—determining services to be offered and rates. The facility applications have been the most studied, tend to be the simplest, and have yielded effective engineering solutions using computer programs for many recurring problems. Operations problems tend to be more difficult both to formulate and to solve, and their large size and complexity have frequently led to approximation in both formulation and solution. Marketing 2 problems usually show complex interactions among the level of demand, the plant and services required to satisfy this demand, and the resulting cost of the services. This area has major difficulties in formulation and solution and has been the least studied of the three. However, the marketing area is likely to become the most important one in the future, primarily because of its basic relation to determining the direction of the business.

Once again we show that the analysis of a very simple artificial example contrived by its authors for a general indictment of all chance constrained programming is itself in error. Proofs and further references are supplied below along with a discussion of other deficiencies and ambiguities in an article authored by A. J. Hogan, J. G. Morris and H. E. Thompson in Management Science, Vol. 27, No. 6 (June 1981).

Using what he calls a “Basic Chance Constrained Programming Model,” BCCM, R. A. Blau proceeds in Blau [Blau, R. A. 1974. Stochastic programming and decision analysis: an apparent dilemma. Management Sci. 21 (3, November) 271–276.] to derive some results which he believes lead to a dilemma in which EPVI, the Expected Value of Perfect Information, is less than EVSI, the Expected Value of Sample Information. This is one part of Blau's dilemma. In addition, EPVI and EVSI are both negative. This is the other part of Blau's dilemma, from which he goes on to a variety of philosophic and other objections that he believes relate to contrasts between Bayesian decision theory and other, more classic, approaches in statistics. The latter are, in turn, related to chance constrained programming, and so on, but in ways that are only adumbrated rather than developed in detail by Blau.

This paper was prepared for presentation on October 15, 1965, as a luncheon address at The Institute of Management Sciences Eastern Meeting held at the Sheraton Hotel, Rochester, New York. Some of our thoughts on this subject were stimulated by a TIMS sponsored methodology symposium on implementation held at Monterey California, during December of 1964 and, which was subsequently followed by our participation in a round table conference organized by Harvey Wagner for the 1965 National Meetings of TIMS in San Francisco on February 3, 1965. Subsequent discussions with G. Kozmetsky of Teledyne, Inc., D. B. Learner of BBDO, Inc., and B. M. Rowles of M&M's Candies, Inc., were also important in helping us to shape the ideas that are presented in this paper.

The July, I960, issue of Management Science contains an English translation of an important original article by L. V. Kantorovich [Kantorovich, L. V. Mathematical Methods of Organizing and Planning Production. Leningrad University, 1939, with a Foreword by A. R. Marchenko (Russian). An English translation, prepared by R. W. Campbell and W. H. Marlow, appears under this same title in Management Science Vol. 6, No. 4, (July, 1960), pp. 366–422.] and an introductory note by T. C. Koopmans [Koopmans, T. C. 1960. A note about Kantorovich's paper, ‘Mathematical Methods of Organizing and Planning Production'. Management Sci. 6 (4, July) 363–365.] which is both evaluative and explanatory. In his penultimate paragraph, Professor Koopmans accords a tribute to this writing of Kantorovich which is at least partly deserved, although it is also paid for, we think, at too high a price in modesty for Professor Koopman's own work as well as the work of others. There is a certain lack of precision at critical points in the Kantorovich article which may tend to compound this price if allowed to stand through time. Professor Koopmans does not take adequate account of this lack of precision—e.g., in his discussion of the computing methods—perhaps because he restricted himself to a very short preface. Also, he runs together various major ideas (resolving multipliers, duality and efficiency prices) which need to be separated for more careful examination. Finally, the discussion is so cryptic as to raise some questions as to what Koopmans had in mind in his closing references wherein he compares Kantorovich's accomplishments (in the instant article) to those of von Neumann, Dantzig, et al. In this paper we shall examine some of these ideas at greater length than Koopmans permitted himself but without necessarily confining ourselves to the instant article by Kantorovich or to its preface by Koopmans.

When a system is described in terms of a linear programming problem max c T x with Ax ≦ b, x ≧ 0, study of its properties, e.g., sensitivity analyses, etc., focuses on effects of alterations in the triple (A, b, c) on the properties of the system. These effects are non-linear and generally lead one away from a model with convenient special structure to much more complex systems. In this paper, methods (“repricing-reprocessing” theorems) are developed which (under certain assumptions about real world behavior) permit one to assess these effects exactly by means of a model of the same structure which can be prescribed in advance. The proofs of these theorems are accomplished by the “chained construction” methods of Charnes and Cooper. Approximation techniques and exact characterization of the non-linearities are also presented.

A new conceptual and analytical vehicle for problems of temporal planning under uncertainty, involving determination of optimal (sequential) stochastic decision rules is defined and illustrated by means of a typical industrial example. The paper presents a method of attack which splits the problem into two non-linear (or linear) programming parts, (i) determining optimal probability distributions, (ii) approximating the optimal distributions as closely as possible by decision rules of prescribed form.