A method for optimally selecting a subset of decisions from a first plurality of decisions related to management of a chaotic event. The first plurality of decisions related to the chaotic event is received. A heuristic algorithm is used to eliminate a first subset of decisions. The first subset of decisions is in the first plurality of decisions. A second plurality of decisions is formed. The second plurality of decisions comprises the first plurality of decisions less the first subset of decisions. A mathematical optimization algorithm is then used to select a second subset of decisions. The second subset of decisions is within the second plurality of decisions. The mathematical optimization algorithm takes as input at least one constraint and chaotic event information. The second subset of decisions is stored.