Thomas R. Sexton

Bio: Thomas R. Sexton is an academic researcher from Stony Brook University. The author has contributed to research in topics: Data envelopment analysis & Inefficiency. The author has an hindex of 21, co-authored 57 publications receiving 2992 citations. Previous affiliations of Thomas R. Sexton include State University of New York System.

FuzzEA: A Fuzzy Logic Approach to Efficiency Analysis

Abstract: We apply fuzzy logic to measure the efficiency of a unit that consumes multiple inputs to produce multiple outputs. We demonstrate how the fuzzy inference process produces an efficiency profile of a producing unit, which can serve as the output of the analysis or which can be “defuzzified” to produce a specific efficiency score. The approach, which we call FuzzEA for Fuzzy Efficiency Analysis, allows the analyst to measure the efficiency of an individual unit without collecting detailed data for all comparable units. The approach allows the analyst, in conjunction with the context expert, to incorporate industry-specific experiential knowledge concerning relevant ratios of inputs and outputs, information often known as benchmarks, or “rules of thumb.” Therefore, the analyst may prefer the FuzzEA approach when detailed data on all producing units are either unavailable or cumbersome to collect, or when the analyst requires efficiency results for only one or a small number of units. Finally, contex...

A model to assess the quality-cost tradeoff in nursing homes.

Abstract: We develop a model for assessing quality-cost tradeoffs as they arise in nurse staffing and task assignment decisions commonly faced by nursing home administrators and industry regulators. As prospective reimbursement schemes and resource utilization groups (RUGs) are implemented, we expect that tradeoffs such as these will become even more critical than they are now. The model permits the inclusion of quality-related considerations, such as restricting tasks to personnel at specified skill levels, requiring minimum staff levels at each skill level, and selective raising of wage rates. The model is also sensitive to changes in patient needs, labor availability, and technological innovations, permitting its use for general staffing purposes. Because it utilizes linear programming techniques, standard computer software can be used to find solutions very quickly, even on a small personal computer. We describe the application of this model to a nursing home in upstate New York, but the procedure is widely applicable to other health care facilities.

Innovative Design for Courses with Group Projects

Abstract: In this paper, we describe an innovative design for courses with group projects that we have implemented in the graduate Decision Support Systems DSS course offered at the Harriman School for Management and Policy, College of Business, at Stony Brook University. Each group is associated with two projects. For one project, the group acts as the client that is interested in purchasing the DSS under development by a second group. For the other project, the group acts as the developer of a DSS that a third group is interested in purchasing. This design enables students to experience the technical and managerial aspects of software development from the perspectives of both the buyer and the seller. In addition, the design increases each student's in-depth experience with two projects rather than one, thereby providing the student with a broader understanding of the course material. We perform an assessment of the design and present results from a questionnaire about student perceptions.

Using Data Analytics to Improve Nursing Home Quality

Abstract: Background: There exists an array of quality performance measures for nursing homes. They can confuse consumers, administrators, and government regulators. Our methodology provides a unified multidimensional evaluation. Objective: To present a methodology to perform a multidimensional assessment of each nursing home within any specified group of nursing homes to aid policy makers, administrators, and consumers with a clear, easy-to-interpret evaluation of a nursing home quality performance. Methods: We use data envelopment analysis (DEA) to integrate several quality measures into a comprehensive benchmarking model. We present statewide results comparing DEA performance scores with the Five-Star rating using data from New York State (NYS) Department of Health. Results: In total, 212 of the 526 nursing homes performed as well as possible. Public nursing homes are most likely to lie on the frontier and have the highest average performance scores. The relationship between the DEA-based performance scores and the NYS Five-Star quality ratings is very weak. Conclusion: DEA is a comprehensive methodology for measuring nursing home quality. The DEA factor performance scores provide detailed information for individual nursing homes, enabling administrators to benchmark their facility's quality performance and to focus quality improvement efforts more effectively.

Marginal Analysis of Perishable Inventory Models Using the Externality Principle

Abstract: The exact analysis of optimal ordering policies for periodic-review perishable inventory systems is challenging because of their high-dimensional state space arising from multiple interrelated ordering decisions over many periods and age distributions of on-hand products. We develop a new marginal analysis framework to approximate this complex multiple-decision model into a single-decision model with an externality term, which internalizes the long-term impact of ordering decisions. Our externality-based approximation utilizes a constant base-stock policy; it is fast and easy to apply. Numerical experiments show that our approach provides state-dependent ordering amounts almost identical to the optimal dynamic programming-based policy.

Algorithms for the vehicle routing and scheduling problems with time window constraints

Abstract: This paper considers the design and analysis of algorithms for vehicle routing and scheduling problems with time window constraints. Given the intrinsic difficulty of this problem class, approximation methods seem to offer the most promise for practical size problems. After describing a variety of heuristics, we conduct an extensive computational study of their performance. The problem set includes routing and scheduling environments that differ in terms of the type of data used to generate the problems, the percentage of customers with time windows, their tightness and positioning, and the scheduling horizon. We found that several heuristics performed well in different problem environments; in particular an insertion-type heuristic consistently gave very good results.