Asymptotic formulas for the derivatives of probability functions and their Monte Carlo estimations

These tools are implemented in penalized optimization routines adapted to stochastic optimization, and are shown to reduce the computational cost of chance constrained programming substantially.