Economic Optimization by Simulation: The Confidence Level Approach

Summary

Deterministic Optimization schemes are of little value to help decision-making in business. One must cope with statistics which show neither mathematical functional forms nor extend over periods of time comparable to the characteristic time periods over which business projections can be averaged. Using factual statistics, one is reduced to Monte Carlo sampling methods which lead to sets of possible chains of events, and corresponding probabilities of occurrence. The introduction a priori of the confidence level one wishes to maintain permits one to assess the results in an easily understood and meaningful manner, reduce machine run-time, take into account statistics introduced for any reason, take into account deterministic and conditional strategies and even sets of possible strategies of opponents. Strategies can be ranked by the confidence with which they can reach or surpass a given goal.