Susan W. Palocsay

Image space analysis of generalized fractional programs

The solution of a particular nonconvex program is usually very dependent on the structure of the problem. In this paper we identify classes of nonconvex problems involving either sums or products of ratios of linear terms which may be treated by analysis in a transformed space. In each class, the image space is defined by a mapping which associates a new variable with each original ratio of linear terms. In the image space, optimization is easy in certain directions, and the overall solution may be realized by sequentially optimizing in these directions.In addition to these ratio problems, we also show how to use image space analysis to treat the subclass of problems whose objective is to optimize a product of linear terms. For each class of nonconvex problems, we present an algorithm that locates global solutions by computing both upper and lower bounds on the solution and then solving a sequence of linear programming sub-problems. We also demonstrate the algorithms described in this paper by solving several example problems.

An optimization model for planning natural gas purchases, transportation, storage and deliverability

Natural gas local distribution companies (LDCs) face the problem of managing natural gas purchases under conditions of uncertain demand and frequent price change. In this paper, we present a stochastic optimization model to solve this problem. Unlike other models, this model explicity considers deliverability, the rate at which gas can be added to and withdrawn from a storage facility, as a variable, and considers its role in ensuring a secure supply of gas. Deliverability is often overlooked in gas supply planning, yet is a critical factor in achieving a secure gas supply. Using data from an LDC in Huntsville, Alabama, we show how this model can be used to minimize total cost while meeting constraints regarding the security of gas supply. We also demonstrate that security is dependent on the rate of deliverability, which in turn is affected by a number of factors including gas availability, storage and transportation considerations, and weather conditions.

Cellular vs process layouts: an analytic investigation of the impact of learning on shop performance

Evidence from the literature on cellular manufacturing suggests that shops configured as manufacturing cells perform poorly compared to job shops. However, cellular shops, being conducive to the use of teams in the assignment of production activities, have the potential to yield higher productivity than a job shop. Productivity differentials, and in particular, differences in the rates at which processing times can be reduced, have been largely overlooked in prior comparisons of cellular and job shops. This paper uses queuing theory to illustrate the relationship between processing time learning rates and flow time performance in cellular and job shops. Models are developed that make it possible to estimate the learning rate required in a cellular shop in order for it to yield performance comparable to that of a job shop. Simulation is used to validate the models under dynamic conditions as opposed to the steady state conditions assumed by queuing theory. Results indicate that a cellular shop need only achieve a marginally higher learning rate than a job shop in order to perform at a comparable level.

AN EXCEL-BASED DECISION SUPPORT SYSTEM FOR SCORING AND RANKING PROPOSED R&D PROJECTS

One of the most challenging aspects of technology management is the selection of research and development (R&D) projects from among a group of proposals. This paper introduces an interactive, user-friendly decision support system for evaluating and ranking R&D projects and demonstrates its application on an example R&D program. It employs the scoring methodology developed by Henriksen and Traynor to provide a practical technique that considers both project merit and project cost in the evaluation process, while explicitly accounting for trade-offs among multiple decision criteria.1 The framework of the Excel-based system, PScore, is presented with an emphasis on the potential benefits of using this methodology with computer-automated extensions that facilitate and enhance managerial review and decision-making capabilities.

Neural Network Modeling in Cross-Cultural Research: A Comparison with Multiple Regression

This article describes the use of neural networks as an alternative method to investigate the links between various dimensions of culture and perceptions of justice and demonstrates their ability to model the data relationships with higher accuracy than multiple regression analysis. A complete discussion of the development and validation of the neural network models is included as a guide to researchers in management who are interested in exploring this methodology.

Predicting criminal recidivism using neural networks

Abstract Prediction of criminal recidivism has been extensively studied in criminology with a variety of statistical models. This article proposes the use of neural network (NN) models to address the problem of splitting the population into two groups — non-recidivists and eventual recidivists — based on a set of predictor variables. The results from an empirical study of the classification capabilities of NN on a well-known recidivism data set are presented and discussed in comparison with logistic regression. Analysis indicates that NN models are competitive with, and may offer some advantages over, traditional statistical models in this domain.

Interdisciplinary Collaborative Learning: Using Decision Analysts to Enhance Undergraduate International Management Education

This article describes an experiential learning activity designed to promote the development of decision-making skills in international management students at the undergraduate level. Students from an undergraduate management science course in decision analysis served as consultants on a case assigned to teams in an international management class. Communication between groups and consultants was facilitated by the use of an Internet-based Web conferencing system. Surveys completed by international management students indicated that the exercise enhanced their understanding of the role of quantitative analysis in an important business decision and increased their confidence in the group’s final recommendation in the case.

An Investigation of U.S. Undergraduate Business School Rankings Using Data Envelopment Analysis with Value-Added Performance Indicators.

Bloomberg Businessweek ranks U.S. undergraduate business programs annually. These rankings provide a convenient overall measure of quality, which is important in todays environment of concern about higher education costs and employment after graduation. Data envelopment analysis (DEA) has advantages over previous regression approaches in characterizing value added. The authors use a DEA approach to estimate relative efficiencies based on starting salaries and recruiter surveys, identifying some schools as overachievers relative to their Bloomberg Businessweek rankings. In particular, DEA-based reranking highlights the ability of some public institutions facing high student–faculty ratios to turn out well-regarded graduates with high starting salaries.

The Complexity of Minimum Ratio Spanning Tree Problems

We examine the complexity of two minimum spanning tree problems with rational objective functions. We show that the Minimum Ratio Spanning Tree problem is NP-hard when the denominator is unrestricted in sign, thereby sharpening a previous complexity result. We then consider an extension of this problem where the objective function is the sum of two linear ratios whose numerators and denominators are strictly positive. This problem is shown to be NP-hard as well. We conclude with some results characterizing sufficient conditions for a globally optimal solution.

An Empirical Evaluation of Probability Estimation with Neural Networks

Recent interest in neural networks by researchers across a wide spectrum of disciplines has provided convincing evidence of their ability to address classification problems. In this article, we consider the issue of evaluating the predictive capability of neural networks when the output values are to be treated as probabilities. We propose the use of a variant of a chi-square statistic, based on the Hosmer–Lemeshow statistic from logistic regression, to measure the goodness-of-fit of neural network models for two-group membership problems. Through experimentation with a large real-world database, we demonstrate the application of this statistic, and examine the effects of varying the number of nodes in the hidden layer on its value. Our empirical results suggest that this statistic can be very useful in identifying significant differences in the probability estimation accuracy of neural network models.