Knowledge Content Library
Expert elicitation: an application at Dow Agrosciences
Saurabh Bansal, Pennsuylvania State University
Motivated by a unique agribusiness setting at Dow AgroSciences, this presentation will discuss an optimization-based approach to estimate the mean and standard deviation of probability distributions from noisy quantile judgments provided by experts. The approach estimates both the mean and standard deviation as weighted linear combinations of quantile judgments, where the weights are explicit functions of the expert's judgmental errors. The approach is analytically tractable, and provides flexibility to elicit any set of quantiles from an expert. It also shows analytically that the weights for the mean add up to one and the weights for the standard deviation add up to zero -- these properties have been observed numerically in the decision analysis literature in the last thirty years, but without a systematic explanation. The approach also establishes that using an expert's quantile judgments to deduce the distribution parameters is equivalent to collecting data with a specific sample size. The theory has been in use at Dow AgroSciences since two years for making an annual decision worth $800 million. The use of the approach has resulted in the following monetary benefits: (i) firm's annual production investment has reduced by 6--7%, and (ii) profit has increased by 2--3%. The last part of the presentation will discuss some practical guidelines for using expert judgment for operational uncertainties in industrial settings, based on the experience at Dow AgroSciences.
Click on the file below to hear a sample of the presentation.
Click here for access to the full video and pdf.
SDP membership is required for full access to this and all other archived webinars.
Keywords: probability assessment probass, agriculture, modeling modtree, risk and uncertainty riskunc, expert interviews expint, optimization optz