Knowledge that Transforms

Every working day between 5 and 6 P.M., 20,000 New Jersey bound commuters enter the Port of New York Authority Bus Terminal in New York City. Over 400 buses arrive in the terminal, load these commuters and depart during this one peak hour. This terminal is located in the heart of the city where land availability is critical. Thus, “limited access” multiple berth platforms are utilized. The commuter operation in the terminal is characterized by the following conditions: 1. Two random inputs cause two converging inter-related waiting lines to be formed—one of buses and one of passengers. Bus waiting lines are critical because of limited bus storage areas. 2. “Limited access” berthing which prevents buses from passing in platform lanes necessitates a commuter traffic simulation method to govern the problem's internal logic. 3. Parameters for bus arrival distribution functions can be controlled to some degree by scheduling policy. The method provides for evaluation of a variety of such policies. 4. Three parties with three independent interests—The Port Authority, the bus owners and operators, and the commuters—must be simultaneously satisfied. Because of these conditions; optimum design constitutes a challenging research problem. A Port Authority Operations Research team studied the problem using traffic simulation and Monte Carlo methods on an IBM 650 Computer.

The purpose of this paper is to illustrate some applications of the functional equation technique of the theory of dynamic programming to a general class of problems arising in the study of networks, particularly those arising in transportation theory.

The “Management Sciences” must have a science foundation upon which to build a structure to serve management. To date, the sciences making the greatest contribution have been mathematical and physical in nature. But theory, experimental design, and measurement are rapidly revolutionizing our knowledge of the human being so that the psychological and biological sciences may be expected to make a greater and greater contribution to the management sciences. This paper reports the work being done on the research frontier at the University of Michigan in connection with organisms as systems, the reactions of groups, models of human preference, the experimental exploration of human observation, detection, and decision making, and the analysis and synthesis of speech. While these areas are not useful for immediate application at the user's level in the management sciences, a recent symposium on the subject provided a short tour through these greenhouses of science for management from which applicable science may be expected to emanate. The material of the symposium will be reported in the paper.

My predecessors, the three former presidents of The Institute of Management Sciences, have already established something of a tradition for the TIMS presidential address. Having been involved with the affairs of the Institute during most of a year, they talked about its aims and activities. This address also will concern these activities; in fact, that is its title: The Institute in Action. Before this address is over I hope you will have reflected on something which you already know and that you will carry this thought with you, if no other, when you leave. You are the Institute and the Institute in Action is you in action.

Most analyses of decision making define decision as choosing from among alternative courses of action. Usually, the criterion of choice is assumed to be some function of probability and utility of the different outcomes associated with each course of action. It is shown here that decision is more realistically concerned with constructive, purposive, or creative aspects of human behavior. In addition to the usual decisions of action, other classes of decision include decisions of understanding, of recognition, and of enterprise. These classes are interdependent and their members interrelated in any realistic situation. Organization planning is shown to be concerned largely with interrelating decisions and with providing criteria for decision. The quality of organization planning affects the quality of decisions, the degree of human satisfaction, and the effectiveness of the organization.

During the past few years, many interesting papers have been written on the subject of product-market diversification. A majority of the writers have dealt with either case histories of successful diversification or with qualitative check-off lists to be used in analyzing specific diversification opportunities. A much smaller group of papers has been devoted to formulation of a systematic approach which a company can use to compare alternative diversification decisions. This paper falls in the latter category. As a first step, diversification is defined and distinguished from other company growth alternatives. Typical growth perspectives are described which may motivate a company to diversify. Diversification objectives are established and related to the company's long-range objectives. A two-step evaluation scheme is proposed for selection of the preferred diversification strategy. The first is a qualitative step, which narrows a wide field of diversification opportunities to a selected few which are consistent with the company's diversification objectives and long-range policy. In the second step, a quantitative procedure is outlined for evaluating the relative profit potential of the selected alternatives. Finally, limitations of the present method are discussed.

The logical structure of diversification as a management policy is discussed through the use of simple examples. For a decision maker who maximizes expectation in the face of uncertain prospects, the kinds of utility functions which lead to the use of diversification are explored. It is shown that conditions of positive but decreasing marginal utility must prevail, and the conditions for optimal diversified programs are given. The use of diversification as a policy for the reduction of variance of outcomes is discussed. For a decision maker whose choices are rationalized in terms of aspiration levels, the situations in which this policy might be used are also explored.

Despite the success of numerous sporadic efforts in the study of organizations, there is at present no unified theory of organizations. This would require a clearcut conceptual apparatus and a methodology for prediction. To achieve the former within the foreseeable future it might be wise to tie into the existing terminology of game theory; the latter may require taking a new look at the psychological problems involved and introducing something like an operations research approach to them.

Consideration is given to problems of choosing the order of issue of items from a stockpile of material whose utility characteristics are changing with time. Conditions are given under which either LIFO (last in, first out) or FIFO (first in, first out) is an optimal issue policy.

Scheduling heating oil production is an important management problem. It is also a complex one. Weather and demand uncertainties, allocation of production between different refineries, joint- and by-product relations, storage limitations, maintenance of minimal supplies and many other factors need to be considered. This paper is concerned with one of an integrated series of operations research studies directed toward improvement in such scheduling methods. Emphasis is on essentials of the mathematical model. Institutional features and other phases of the OR studies are brought in only as required.