Ronald K. Klimberg

Modeling data envelopment analysis (DEA) efficient location/allocation decisions

Many types of facility location/allocation models have been developed to find optimal spatial patterns with respect to various location criteria that include cost, time, coverage, and access among others. In this paper we develop and test location modeling formulations that utilize data envelopment analysis (DEA) efficiency measures to find optimal and efficient facility location/allocation patterns. We believe that solving for the DEA efficiency measure, simultaneously with other location modeling objectives, provides a promising rich approach to multiobjective location problems.

Benchmarking Academic Programs in Business Analytics

Interest in business analytics (BA) is currently popular. Professional consultancies and software houses are both touting it as the next wave in business, claiming that the need for BA skills is large and growing. Universities are beginning to respond by offering undergraduate majors and minors, Master of Science degrees, certificates, and concentrations within their Master of Business Administration programs. But what subjects are being covered in these programs? We surveyed some of the largest, most established, and best-known programs (predominantly in the United States, but some international) and interviewed representatives of these programs to better understand the requirements for students entering, the required and elective course topics covered, and job opportunities for graduates. In this article, we summarize our findings and provide some conclusions about analytics programs, including the current landscape, suggestions for development, and our vision for the future. We believe this report is useful to institutions that offer analytics programs, to those considering such offerings, and to the employers who are hiring analytics professionals. These employers need to better understand the skills that professionals are acquiring. Finally, it should help prospective students who seek to understand the analytics programs being offered to find the best match for their skills and interests.

Business Plus Intelligence Plus Technology Equals Business Intelligence

In this article the authors will show how the parallel developments of information technology at the operational business level and decision support concepts progressed through the decades of the twentieth century with only minimal success at strategic application. They will posit that the twin technological developments of the world-wide-web and very inexpensive mass storage provided the environment to facilitate the convergence of business operations and decision support into the strategic application of business intelligence.

AGGREGATE PLANNING MODELS FOR FIELD SERVICE DELIVERY.

Abstract A major planning problem for any after-sales repair service operation concerns the siting and staffing of field offices. Field service management seeks to balance costs against customer satisfaction, which is usually measured in responsiveness. This paper applies set covering models to the problem of field service aggregate planning. Four model formulations are developed and tested with a subset of actual data. The results of these aggregate planning models indicate which offices should be opened and how many field engineers of various expertise levels should be placed in each office to meet the dual objectives of minimizing cost and ensuring adequate responsiveness. Three of the four models are multiple objective formulations that demonstrate the tradeoffs between cost and responsiveness for varying levels of staffing and office openings.

Experimental evaluation of a graphical display system to visualizing multiple criteria solutions

Abstract One of the major limitations of multiple criteria techniques is the poor set of mechanisms available for evaluating alternatives/solutions. Simply listing the alternatives, in most cases, is not satisfactory. Alternative solutions must be communicated to the decision maker in a clear and concise manner that makes it easy for the decision maker to explore and evaluate their respective tradeoffs. This paper describes a graphical display system (GRADS) that presents multiobjective solutions. GRADS allows the decision maker to explore the tradeoffs between objectives from a given set of finite alternative solutions. The paper presents the results of an experiment that tested the usefulness of GRADS in the problem-solving process. Additionally, a relatively new statistic called Hildebrands del is used to provide further insight into the experiments results.

Forecasting performance measures – what are their practical meaning?

Producing good forecast is a vital aspect of a business. The accuracy of these forecasts could have a critical impact on the organization. We introduce a new, practical, and meaningful forecast performance measure called percentage forecast error (PFE). The results of comparing and evaluating this new measure to traditional forecasting performance measures under several different simulation scenarios are presented in this chapter.

Experimental comparison of parametric, non-parametric, and hybrid multigroup classification

Highlights? The proposed hybrid method dominates almost all the other methods on classification performance. ? Logistic regression and neural network provides worst relative performance under most scenarios. ? This shows that the data complexities have adverse impact on the multinomial logistic regression. ? The results indicate that all classification methods are adversely affected by the nonstatic data. ? This study demonstrates the effectiveness of the hybrid method in improving classification accuracy. This study evaluates the relative performance of some well-known classification techniques, as well as a proposed hybrid method. The proposed hybrid method is a combination of k-nearest neighbor (kNN) and linear programming (LP) method for four group classification. Computational experiments are conducted to evaluate the performances of these classification techniques. Monte Carlo simulation is used to generate dataset with varying characteristics such as multicollinearity, nonlinearity, etc. for the experiments. The experimental results indicate that LP approaches, in general, and the proposed hybrid method, in particular, consistently have lower misclassification rates for most data characteristics. Furthermore, the hybrid method utilizes the strengths of both methods - k-NN and linear programming - resulting in considerable improvement in the classification accuracy. The results of this study can aid in the design of various hybrid techniques that combine the strengths of different methods to improve classification accuracy and reliability.

Enterprise Information System and Data Mining

The advent of information technology and the consequent proliferation of information systems have lead to generation of vast amounts of data, both within the organization and across its supply chain. Enterprise information systems (EIS) have added to organizational complexity, and at the same time, created opportunities for enhancing its competitive advantage by utilizing this data for business intelligence purposes. Various data mining tools have been used to gain a competitive edge through these large data bases. In this paper, the authors discuss EIS-aided business intelligence and data mining as applicable to organizational functions, such as supply chain management (SCM), marketing, and customer relationship management (CRM) in the context of EIS.

Creating an index of vulnerability to severe coastal storms along the North Shore of Boston

A great deal of uncertainty accompanies predictions of the potential effects of global climate change on the coastal hazards associated with severe storms. One way to obviate the effects of this uncertainty on the design of policies is to understand the manner in which populations are currently vulnerable to these types of hazards. In this chapter, we develop a method for constructing a relative composite measure of vulnerability using data envelopment analysis (DEA). Through the application of this index, and one constructed using a weighted average, to four costal towns along Bostons North Shore, we demonstrate their potential usefulness to policy formulation and implementation. The DEA composite index is shown to complement the information provided by the weighted average and helps overcome some of its shortcomings such as assigning importance weights and masking of the influence of one or a subset of vulnerability attributes. Acknowledging the spatial implications of floodplain protection and mitigation efforts, the indices are constructed and analyzed at a number of different geographic scales.

Data-Driven Decision Making for New Drugs: A Collaborative Learning Experience

Two courses, advanced decision-making and pharmaceutical marketing, were combined in a collaborative process to mimic how the pharmaceutical industry determines the potential of new drugs. Integrated student teams worked together to complete semester-long projects and taught each other their respective knowledge areas—marketing and statistics. Real-world data for medical and pharmacy claims payments were “cleaned†and mined by students to analyze usage and cost patterns for anti-hypertensive and anti-hypercholesterolemia drugs currently on the market. Analyses included merging the medical and pharmaceutical data records to derive individual electronic patient records, which were the basis of financial projections for the new drugs. Importantly, the single patient record is congruent with the needs of the stakeholders currently working to reform U.S. healthcare delivery.