Industrial engineering's Johnson to give talk Monday

Dr. Andrew Johnson, assistant professor in the Department of Industrial and Systems Engineering at Texas A&M University, will give a talk Monday (May 3) at 3 p.m. in Room 203 of the Zachry Engineering Center on campus.Dr. Andrew JohnsonJohnson's lecture, "Nonparametric Efficiency Estimation Methods in the Presence of Different Operating Environments: A Warehouse Industry Benchmarking Analysis," is part of the Department of Industrial and Systems Engineering's seminar series, sponsored by Parsons.Abstract Understanding the effects of operational conditions and practices on productive performance can provide insight into underlying causes of inefficiency and indicate better managerial strategies. To this end, a large number of studies have employed a two-stage data envelopment analysis (DEA) method where the DEA efficiency estimates are regressed on contextual variables representing the operational conditions. Unfortunately, the two-stage DEA estimator is biased when the contextual variables are correlated with inputs. To circumvent this problem, we build upon the results of Kuosmanen and Johnson (2010) who showed Convex Nonparametric Least Squares (CNLS) regression might be used to estimate DEA efficiencies.Using this bridge between the DEA and regression paradigms, we develop two new one-stage estimation approaches. The first approach called one-stage DEA directly integrates the contextual variables to the standard DEA model. The second approach applies CNLS regression with contextual variables to estimate the shape of the frontier, and subsequently estimates the efficiency scores based on CNLS residuals. The latter approach allows for correlation between the contextual variables and the input levels. Monte Carlo simulations demonstrate that proposed estimators perform consistently better than two-stage DEA in a wide variety of scenarios.To demonstrate the use of these methods for practical problems, we analyze the warehousing industry. Warehouses are a substantial component of logistic operations, and a significant contributor to speed and cost in supply chains. While there are widely accepted benchmarks for individual warehouse functions, such as order picking, there is less general knowledge of the overall technical efficiency of warehouses. The lack of a general understanding of warehouse technical efficiency and the associated causal factors limits the ability to identify the best opportunities for improving warehouse performance. While there have been efforts to analyze warehouse efficiency, there is still a significant gap in the education and training of the industry's professionals. The second part of this talk will discuss the application of one-stage DEA and CNLS regression with contextual variables to a large sample of warehouses providing self-reported attribute and performance data. The results are contrasted with the results given by the standard two-stage DEA estimator.Biography Dr. Andrew L. Johnson is an assistant professor in the Department of Industrial and Systems Engineering at Texas A&M University. He obtained his B.S. in industrial and systems engineering from Virginia Tech and his M.S. and Ph.D. from the H. Milton Stewart School of Industrial and Systems Engineering from Georgia Tech. His research interests include productivity and efficiency measurement, warehouse design and operations, material handling and mechanism design. He is a member of the INFORMS, National Eagle Scout Association, and German Club of Virginia Tech.