The Deterministic Equivalents of Chance-Constrained Programming

Three concepts combine to show both the feasibility and desirability of incorporating probability within programming models. First, the reliability of estimates obtained by using Chebyshev's inequality increases as variation measured by the coefficient of variation, declines. Second, the coefficient of variation can be substantially reduced by the use of the mean and variance of a truncated normal distribution. Third, chance-constrained programming can be converted into deterministic equivalent quadratic programming by using the parameters of a truncated normal distribution.