Planning for Hurricane Isaac using Probability Theory in a Linear Programming Model

“

" The most important point of the article was the risk analysis and planning in the second scenario. The second LP model demonstrated that uncertainty could be included as calculated risk. Model 2 illustrated how uncertainty was quantified as risk by calculating the means, standard deviations and coefficient of variations for airboat trips based on historical data from hurricane Katrina. Risks must be estimated and configured carefully in LP models. The objective function and decision variables along with a minimization or maximization goal will drive the search for a feasible solution in an LP model. This means that the objective function coefficients (costs or revenues) have the most impact on the decision variables. Therefore, if there are large numerical differences between the objective function coefficients, this will greatly influence the resulting decision variables calculated by Solver, notwithstanding the coefficients. Other researchers can experiment with this concept by coding a dummy constraint line and trying different objective function coefficients with a minimization or maximization goal. In fact, this is the recommended approach for setting up a new LP model. In this way a researcher will better understand their LP model rather being surprised by irrational (non-logical) results from Solver." (p. 63)

”