Analysis of the Sensitivity of Decision Analysis Results to Errors and Simplifications in Problem Structure: Application to Lake Erie Ecosystem Management

In practical decision analyses, the ldquocurse of dimensionalityrdquo compels one to make simplifying assumptions that can introduce errors into estimates of various indexes that interest decision-makers. These indexes include the expected performance of optimal and suboptimal strategies, the benefit of explicitly considering uncertainty, and the benefit of additional information. This paper quantifies the effects on these indexes of simplifying assumptions, including discretization of the decision space, omission of some decision variables and uncertainties from the decision tree, and disregarding of risk aversion. To reduce errors arising from discretization of the decision space, we use a multidimensional cubic spline to interpolate the performance of alternatives between a few simulated points. A case study analyzes decisions concerning phosphorus loading, fisheries management, and lower trophic research projects in Lake Erie under multiple criteria and ecological uncertainties. Results show that spline-based solutions often yield potentially superior decisions from those based on discretized decision spaces, but that omitting important uncertainties makes more of a difference in this case study's decisions and indexes than simplifying the decision space. On the other hand, incorrect multicriteria weights affect the case study's outcomes more than incorrect probabilities.