Increasing manufacturing flexibility is a key strategy for efficiently improving market responsiveness in the face of uncertain future product demand. Process flexibility results from being able to build different types of products in the same plant or production facility at the same time. In Part I of this paper, we develop several principles on the benefits of process flexibility. These principles are that 1) limited flexibility (i.e., each plant builds only a few products), configured in the right way, yields most of the benefits of total flexibility (i.e., each plant builds all products) and 2) limited flexibility has the greatest benefits when configured to chain products and plants together to the greatest extent possible. In Part II, we provide analytic support and justification for these principles. Based on a planning model for assigning production to plants, we demonstrate that, for realistic assumptions on demand uncertainty, limited flexibility configurations (i.e., how products are assigned to plants) have sales benefits that are approximately equivalent to those for total flexibility. Furthermore, from this analysis we develop a simple measure for the flexibility in a given product-plant configuration. Such a measure is desirable because of the complexity of computing expected sales for a given configuration. The measure is ∏(M*), the maximal probability over all groupings or sets of products (M) that there will be unfilled demand for a set of products while simultaneously there is excess capacity at plants building other products. This measure is easily computed and can be used to guide the search for good limited flexibility configurations.

Cognitive styles measured by the Myers-Briggs indicator were studied to isolate how style influences decision behavior. An experiment was conducted in which experienced decision makers from hospitals and firms were asked to assess several capital expansion projects. To control for environmental factors, the project summaries were tailored to be compatible or incompatible with each individual's cognitive style. Risk and information sources were also controlled in the summary by using two levels of risk, defined by the spread of the return on investment (ROI) projections, and by using personal judgements or a computer-based model to provide the ROI estimates. The decision makers assessed each project, indicating their likelihood of adopting it and their perception of its risk. The impact of style, setting (hospital or firm), and the control factors (environment, information source, and risk) were related to a decision to adopt and assessments of risk by ANOVA techniques. Cognitive style was found to be an important factor in the decision to adopt and the assessment of risk. ST (sensation-thinking) styles saw the highest risk and were reluctant to adopt the projects, while SF (sensation-feeling) styles were risk tolerant and more likely to adopt the same projects. Our results support the views of cognitive theorists, who argue that decision style is an important determinant of behavior. Decisions seem to be a function of the decision maker's cognitive makeup which differs for different psychological types.

This paper derives the simple analytic solution to the special but important inventory problem in which the optimal policy is to reorder whenever units are demanded. The demand distribution can be any compound Poisson; the resupply distribution is arbitrary. Both the backorder case and the lost sales case are solved by generalizing a queueing theorem due to Palm. The steady state probabilities for the number of units in resupply (or repair) completely describe the item's long term behavior, and are simply the normalized values of the compound Poisson demand distribution based on the mean of the resupply distribution but not on the distribution itself. Knowledge of these state probabilities enables us to compute several measures of item supply performance as a function of the spare stock, s. Traditional inventory analysis can then be applied to minimize total cost based on estimates of holding cost and supply performance cost. The appendices contain a description of the algorithm and the computer program for calculating stuttering Poisson state probabilities and the measures of effectiveness for the backorder case. Numerical illustrations are also provided.

This paper characterizes optimal service policies in terms of the frequency and timing of services which are intended to maintain a stock of assets. The model is non-stochastic. The results are obtained by a two-stage step-wise minimization procedure wherein dynamic programming is first used to characterize sub-optimal policies and then the calculus of finite differences is utilized to select the optimal policy. The effect of the time horizon on the optimal policy is emphasized throughout and, in two specific interpretations of the model, the classical results of the economic lot-size and equipment replacement policy are shown to be limits of more general policies.

This paper presents a mathematical model for determining a minimum cost strategy for the following personnel recruiting and selection situation. A quota is set for the minimum number of good employees to be added each period, and selection is done on the basis of a test having a known correlation with the given criterion of job success. There is a total cost for all individuals recruited and a loss attached to each person hired who turns out to be unsatisfactory. The problem is to determine a minimum acceptable score on the selection test and the number of persons to be recruited which will yield a minimum cost, including loss, as well as satisfy the prescribed quota at a given probability level.

This paper presents a series of simple but realistic examples in which common algebraic indexing conventions are not so convenient. In particular, it analyzes the difficulties caused by the conventions that each model component must have a fixed number of indices and that the order of the indices is significant to their meaning. To deal with these difficulties compensating extensions to algebraic notation are proposed. The proposed notation is compared to existing notation in terms of the human abilities to understand, maintain and verify model descriptions.

Consumer incentives often have a dual role: to expand the market and to redistribute the market so as to meet sales targets for a range of products. The proper allocation of incentives must recognize and deal with the competitive and “cannibalizing” effects operating among product classes. The presence of environmental factors, scale economies, and cost per unit for each of the candidate incentive types also must be accounted for. This paper develops and illustrates an analytical approach designed to improve the allocation of consumer incentives where sales targets are exogenously set. The model generates a system of simultaneous behavioral equations, a portion of which is included to reflect attempted cost minimizing behavior. The model is applied to FY81-FY86 data from the U.S. Army Recruiting Command, which allocated over $1 billion for special incentives to selected recruits. The incentives consist of enlistment bonuses and educational benefits offered to quality recruits who select critical occupational specialties. The sales targets are enlistment quotas for each of the occupational specialties that must be simultaneously satisfied. The Army's goal is to meet these quotas, in a postulated recruiting environment, at minimum total incentive expenditure. Results suggest that the Army could improve the cost-effectiveness of its recruiting effort through a reallocation of the incentives it offers to recruits. Misallocation during the FY81-FY86 period is estimated to have cost the Army approximately 3.5% of its total incentives expended in infantry recruiting recruiting.

The first major experimental comparison of approval voting with regular plurality voting occurred in the 1985 annual election of The Institute of Management Sciences (TIMS). In approval voting a person votes for (approves of) as many candidates as desired, the winner being the candidate with the most votes. By permitting more votes than the number of positions to be filled, approval voting collects more information from the voter than does plurality voting. This can make a difference, for example, when three candidates compete for a single office. In such situations two candidates with wide but similar appeal sometimes split a majority constituency so that, under plurality voting, a minority candidate is elected. By contrast, approval voting is likely to identify the candidate who is most broadly acceptable to the electorate as a whole. In the TIMS experiment society members received an experimental approval ballot along with their official plurality ballot. Two contests involved three candidates running for a single office and a third, five candidates for two positions. Surprisingly, in two of the three contests, approval voting would have produced different winners and neither of the changes was of the type usually emphasized in the approval voting literature. The experiment demonstrated the practicality of approval voting and showed that it can elect a set of candidates different from that which plurality voting would. Direct comparison of ballots makes it possible to determine why the experimental switches occurred. It is shown that in each reversal the approval winner had broader support in the electorate than the plurality winner. The experiment also provided empirical data on how voters distribute approvals across candidates and indicated that, in this case, their behavior was roughly, but not exactly, consistent with theoretical analyses of voting efficacy.

The standard method for establishing the comparative statics of risk changes in optimization problems has been confined to comparing unique interior solutions, relying on strong assumptions about payoff functions and decision variables. We propose a simple and intuitive approach that hinges on considerably weaker assumptions. Merging insights from the monotone comparative statics literature with insights from the risk apportionment literature, we show that the ranking of simple lottery pairs is all that is needed for establishing the comparative statics of risk changes. We use this approach to analyze the comparative statics of Nth-degree stochastic dominance shifts in a general setting with one and with multiple decision variables, and we show how these results can be applied to generalize the classical theories of precautionary saving, self-protection, and others. This paper was accepted by James Smith, decision analysis.

Expert judgment elicitation is often required in probabilistic decision making and the evaluation of risk. One measure of the quality of probability distributions given by experts is calibration–the faithfulness of the probabilities in an empirically verifiable sense. A method of measuring calibration for continuous probability distributions is presented here. A discussion of the impact of using linear rules for combining such judgments is given and an empirical demonstration is given using data collected from experts participating in a large-scale risk study. It is shown by theoretical argument that combining well-calibrated distributions of individual experts using linear rules can only result in reducing calibration. In contrast, it is demonstrated, both by example and empirically, that an equally weighted linear combination of experts who tend to be “overconfident” can produce distributions that are better calibrated than the experts’ individual distributions. Using data from training exercises, it is shown that the improvement in calibration is rapid as the number of experts is increased from one to five or six, but there is only modest improvement from increasing the number of experts beyond that point.