Search

A study was conducted to analyze the development of an acquisition policy for a single item multi-supplier system. Optimal solution characteristics were correlated with the use of a single source promoted by just-in-time. A single item Economic Order Quantity framework which supported cost parameters and a finite long-run average capacity was utilized. The acquisition policy for a single item was required to meet constraints related to capacity and minimized costs. Some of the costs considered include total periodic purchasing, inventory carrying, ordering, inventory carrying, and supplier management costs. Results indicated that only one supplier can be selected if no capacity constraints existed. In addition, findings showed at least one supplier to supply less than its capacity.

A meta-analysis was performed to find out if research methods had an influence on the results of studies supporting the claim that firms making an early entry into markets have better performance than others that follow their lead. Based on the analysis of 90 statistical tests, it was revealed that tests employing market share as the performance measure had a higher probability of arriving at a first-mover advantage than tests that rely on other indicators, including profitability and survival. Tests that sample from individually chosen industries and those that did not measure the competitive position of the entrants were also found to have a greater likelihood of finding first-mover advantage. The meta-analysis suggests that the first-mover advantage will be less frequent if certain 'limiting' techniques were not applied.

Companies may prevent competition from entering the market by manipulating their prices. For instance, the market share of a firm may be eroded if its prices are high enough to encourage other firms to enter the market. On the other hand, the dominant company's low prices can discourage potential entrants. The strategic pricing behavior of dominant firms operating in multiple markets, particularly the combination of their cost signaling effort across markets, is analyzed.

FW/SM is a research prototype structured modeling (SM) implementation that has been evolving alongside other SM components, such as theory, language and vision. This new implementation has proven its usefulness in several areas, including facilitating the exploration of SM concepts, the study of implementation matters related to modeling environment designs, and the examination of the concept of the ideal model environment. The properties of a modeling environment that will facilitate improved quality and productivity, as well as the design implications of these properties, are discussed.

A review of research on the use of judgment in statistical forecasting indicates that there is validity in the use of judgment in forecasting. The use of judgment is valid if the judgmental process is made explicit by using a type of decomposition or audit trail. The judgmental factors in forecasting include the selection of variables and the analysis of data. The types of research that have helped to substantiate the more critical view of the use of judgment in forecasting include bootstrapping research and quality-of-judgment research. Judgmental forecasters should establish the nature of the comparison, the experiment, and the methods.

An analysis of business literature suggested that there were differences in the performance of US and Japanese firms in terms of software development, and the differences were evaluated by comparing 24 US and 16 Japanese software development projects. The results indicated that the software development practices of Japanese firms were no longer inferior to the practices of US firms. Japanese and US firms developed similar products, and they used similar languages, techniques, and hardware platforms. Japanese projects performed at least as well as US projects in several areas, including productivity, quality, and reuse of software code.

A model of firm performance was developed and tested. The model described the effects of industry structure, business strategy, and managerial characteristics on the performance of new firms. The model was tested by using a sample of 36 new technology-oriented firms funded by venture capitalists. The results indicated that managerial characteristics strongly affected performance, but they did not significantly affect either industry structure or business strategy.

The effect of top management on productivity was investigated by comparing US and Japanese automobile manufacturers. The data were obtained from annual reports of three US and three Japanese automobile manufacturing companies between the early 1950s and 1987. The results indicated that more efficient use of labor was primarily responsible for productivity improvements in all six firms. Management effects were the main source of productivity variations among US and Japanese automobile manufacturers.

A firm's success depends largely on its ability to produce innovative services or products, and firms that create truly innovative new services or products creating new markets or industries are referred to as pioneers. A model of an industry in which a pioneer accumulates capital for a significant amount of time before a new competitor enters the market was constructed to analyze the effect of the order of entry on the market share position of both the pioneer and the competitor. Research results reveal that market share is not dependent on the offer of entry or the period in which the pioneer enjoyed a monopoly. Advantages in goodwill, price, brand loyalty, and lower costs of production and advertising enjoyed by the pioneer dissipate over time and do not affect final market share, nor does the time it took the competitor to overcome them.

Decision makers are faced with the need to incorporate more information into the decision making process as more and better data becomes available. A combination of simple database models with managerial intuition, in which the two are viewed as complementary rather than competitive, is the best method for improving decision quality. A 50% model plus 50% manager heuristic improves the quality of forecasting as statistical models offer more predictive accuracy and managerial intuition provides additional insight models cannot incorporate. The 50% model plus 50% manager heuristic offers the advantages of simplicity, palatability, and accuracy and can be easily incorporated into extant decision support systems.

The evolution of productivity performance at an international electronics equipment manufacturer is analyzed in order to elucidate the cognitive and behavioral processes that are the root of the differences in the total productivity learning curve. Research was based on a field investigation into the cognitive and behavioral determinants of the considerable variation in learning curve parameters across departments. Research results reveal that improvement in manufacturing productivity is dependent on three kinds of shared learning: sharing characterizing the development/manufacturing interface; sharing between the primary plant and plants that come online at a later time; and the continuous sharing that goes on among plants after the start-up.

Non-Markovian stochastic processes can be modeled as Markov Chains when additional components in the state description are included so that the Markovian property is satisfied by the revised state variable's transition. However, this technique leads to a large increase in the number of states, typically to the point where it becomes impractical to apply the model due to the state space dimension. Applying an approximation of the original state variable in condensed form to a specific form of vector-valued Markov chain can address the problem. An intensive algorithm which approximates the conditional distribution for S(t) given X(t) is presented.

Various production-related methods used by firms for gaining entry to foreign markets are analyzed to determine the profits firms earn by using the strategies. The entry techniques analyzed include exclusive licensing, foreign direct investment, and joint venture. Results indicate that an entering firm may be able to alter licensing charges, but the best and most popular profit-generating method is foreign direct investment.

There are many materials for which the quantity needed by a firm is at best indirectly related to the quantity of final product produced by that firm, such as solvents in manufacturing processes or office supplies. For any such 'indirect' materials, an inescapable incentive conflict exists: The buyer wishes to minimize consumption of these indirect materials, while the supplier's profits depend on increasing volume. Both buyer and supplier can exert effort to reduce consumption, hence making the overall supply chain more efficient. However, no supplier will voluntarily participate unless contract terms are fundamentally revised. This can be done through a variety of 'shared-savings' contracts, where both parties profit from a consumption reduction. This paper analyzes several such contracts currently in use for chemicals purchasing. We show that such contracts can always increase supply-chain profits but need not lead to reduced consumption. We analyze equilibrium effort levels, consumption, and total profits, and show how these change with the contract parameters. We find that the goals of maximizing joint profits and minimizing consumption are generally not aligned. Also, surprisingly, a decrease in a cost parameter can lead to a decrease in profits; it may be necessary (but is always possible) to renegotiate the shared-savings contract to reap the benefits of a cost decrease., 1. Introduction The potential for improved supply-chain performance through various forms of coordination has been demonstrated extensively in recent literature and management practice. A wide range of supply contracts has [...]

1. Introduction and Scope This paper is a review of research design and development. We define product development as the transformation of a market opportunity and a set of assumptions [...]

One-way travel time can be improved by dividing warehouses into relatively small number of different regions using a class-based automated storage/retrieval system (AS/RS). Optimal boundaries for the AS/RS can be determined through a one-dimensional solution procedure which allows for the accurate delineation of region boundaries for an AS/RS warehouse.

Much of the research into the growth of organizations focuses on a concept of stages of growth within the organization life cycle and offers multi-stage models based on predictable patterns of growth. Research based on a longitudinal study of 71 computer and electronic production ventures offers a growth model for technology based new ventures (TBNVs) that postulates that TBNVs move through a four stage growth process: conception and development; commercialization; growth; and stability. The research indicates that the four-stage growth hypothesis is valid, although companies display variation in their transition patterns.

Simulation studies of real time control policies can use a modified likelihood-based method to obtain the derivatives from the input parameters of a Poisson process rate. The modification to the likelihood estimator is to use the Poisson rate estimated value during the sample interval. This produces a consistent, asymptomatically normal and unbiased estimate of the derivative for regenerative systems. Mathematical models applying likelihood derivative procedures can be used to optimize the real time control of computer, production, and telecommunication systems.

O.L. Deutsch posed a traveling-salesman problem as a zero-one linear program for the Artificial Intelligence Design Challenge for the 1987 American Institute of Aeronautics and Astronautics. When a nonlinear constraint of the problem is ignored or approximated by linearization, the linear programming problem posed by Deutsch can be solved. A software system, AIAA/SOLVER, has been developed that can solve the problem. The system is based on an assumption about the problem's stochastic cost structure using a branch-and-cut approach. The branch-and-cut approach consists of subroutines that maintain and update data structures representing a search tree. The subroutines are able to comprehend the branches of the search tree and choose a variable on which to branch next.

Importance sampling is an effective technique for increasing the efficiency of Monte Carlo algorithms for the numerical analysis of integrals. The original random mechanism in the simulation can be replaced in order to solve problems arising from the simulation of stochastic systems for application to GI/G/1 queueing problems, including: discrete time Markov chains; continuous-time Markov chains; and generalized semi-Markov processes.

The estimation of the required length of the run of a simulation is a major problem in the planning of simulation experiments addressing the estimation of steady-state quantities related to queueing models. Simple heuristic formulas for one-run queueing simulations can be applied to the design of queueing experiments to aid in approximating queue lengths. The formulas are based on queues with heavy traffic limits and can be applied to stochastic processes reflecting Brownian motion, including the queue process in a standard GI/G/1 model. The formulas can be used to approximate the length of the simulation run in each application of an experiment.

The conditional expectation (CE) variance reduction technique is well-known but unpopular due to its model-dependency and the lack of a general method to find a random variable Y for all computed variables of Y. A modified version of the CE technique, the Partial Conditional Expectation (PCE), provides the capacity to obtain a consistent estimator for computing the waiting time in queue for a single-server queueing system in simulations of a GI/G/1 queue. An improved version of the estimator can be derived by further consideration of the system. The two estimators preform better than the classical estimator and variance reduction obtained from the two estimators is particularly significant with low traffic.

Simulation is valuable to operations research in that it makes it economically feasible to analyze large-scale systems. Variance reduction techniques have been created to increase the efficiency of simulations. The use of control variates, by focusing on linear correlations between output variates, or controls, and target responses, allows the increased efficiency of simulations without manipulating or distorting random number input. Control variates can be used to build a simulation meta-model, applying multiple controls to the meta-models in order to make input, output, and control vectors multidimensional so that the model can be applied to multipopulation, multiresponse experiments.

The estimation of discounted costs can be derived by simulation methodologies for systems displaying stochastic fluctuations in which conventional numerical methods are inefficient. Simulation methodologies utilizing algorithms for discounted cost problems provide a technique for reducing variance when modelling stochastic simulations, including non-Markov processes or infinite state space Markov chains.

Stratified sampling is a natural technique for variance reduction but the high dimensionality of sample space often skews its results. The use of shadow response variables, those variables that have similar properties of the control variables, can be used to significantly reduce sample space variation to within a reasonable tolerance. The selection of good shadow response variables is necessary for successful stratified sampling.

A conceptual flaw that is present in a recent publication is identified regarding the minimization of means squared deviation of completion times about a common due date in manufacturing. Additionally, a modification to methodology is developed which can correct the conceptual flow, as well as improve bounds to determine the nature of a given problem.

Recent analysis is presented which examines how financial decision models have been designed to estimate skewness in stock returns and many financial variables, but ignore the sampling error in skewness estimations. The analysis reveals that there are significant consequences to ignoring skewness sampling error. Additionally, a simple approach is developed which roughly constructs a confidence interval for skewness estimated from lognormal populations.

Research developing a tolerance approach to sensitivity analysis of matrix coefficients in linear programming is described. The tolerance approach to standard linear programming considers simultaneous perturbations in more than one row or column of a linear program and allows for simultaneous and independent perturbations in the matrix coefficients. The concept of approximate tolerance is used to correct for the difficulty in computing the inverse of the perturbed basis when delineating the maximum tolerance under general perturbations. A decision maker can place a priori limits on the possible values over which certain parameters may vary. The tolerance approach delineates a maximum tolerance percentage within which all the elements in a row or column can vary simultaneously and independently from estimated value but still retain the same set of basic variables in an optimal solution.

Univariate forecasting methods are used in production control and are essential to modern manufacturing methods The choice of a forecasting method is critical to production control, particularly to manufacturers with large numbers of production processes and product lines. There are two forecasting approaches: aggregate selection and individual selection. Aggregate forecasting applies a single method to the entire series of production processes. Individual forecasting determines the appropriate method and applies it to each production process, and is an expensive, time consuming process characterized by complex computations. An analysis of 263 data series from a single organization reveals that individual forecasting is a more accurate method for forecasting production processes with long lead times and sample variability, but does not always give superior results due to the lack of stability in data series variables. Additionally, results reveal that aggregate forecasting is a cost-effective compromise that, when applied across a wide cross-section of series and times, is simpler and of comparable accuracy to individual forecasting.

The design of the distribution channel involves the choice of the proper channel format and the appropriate selection of the proper intermediaries. The article examines a decision format and a model that will help management with the tactical selection of markets and channel intermediaries. A real industrial marketing application is presented to support the model-based decision technique.

The effect of advertising on new telephone diffusion in West Germany is examined. Alternate ways of integrating advertising with the 'Bass-diffusion' model are offered and tested empirically. Econometric investigation provides results consistent with a behavioral assumption: advertising mainly affects the demand of imitators. The lag structure entailed in advertising response is best measured by a 'Nerlove-Arrow' model. The maximum effect of advertising is reached several months after it is presented. Seventy-two percent of total advertising effect is shown to occur between four and nine months after advertisements appear.

Free entry can cause excessive competition and too much corporate investment by both successful and unsuccessful corporations. The success-breeds-success assumption is basically responsible for this effect. A three-part policy of patent life restrictions, firm subsidies, and entry taxes can be used to correct such a market failure. One disadvantage of patent life restrictions is that winners may overcompensate and therefore not invest sufficiently, suggesting the need for a consecutive win subsidy.

Three energy options are compared by means of a value tree that relates general values and concerns to certain value-relevant attributes: coal, nuclear, and a package combining geothermal and conservation measures. Judgements on the relative significance of the tree attributes are provided by 37 non-expert subjects. Three basic experiment findings are: (1) hierarchical weights are more steep than nonhierarchical weights, (2) groups that show a different response for holistic first choices nonetheless show significant agreement in assessing attribute weights, and (3) attribute level ratings concerning the relative desirability of energy options are consistent across groups, but there is a tendency for each group to favor its holistic first choice.

Standard scoring rules demand knowledge of an expert's utility function. Basic reference lottery tickets can be employed to develop scoring rules which do not call for such information. Only two distinct values are possible for the final monetary reward when utility is 'linearized' by means of the basic reference lottery tickets. Utility is then a linear payoff function in terms of high reward probabilities. Expert assessments can therefore be gained even when their utility functions are not known or cannot be observed.

A study was conducted to analyze a unique theoretical and empirical approach to tasks such as knowledge management and business process redesign. Examples on how different organizations carry out similar processes were examined and compiled in an on-line process handbook. The handbook was designed to aid people redesign their existing organizational procedures and develop novel organizational processes. It also allows individuals to share their respective ideas about organizational practices. Results indicated the effectiveness of the approach in representing similarities and differences among related processes and its ability to produce sensible options for how a given process could be carried out.