Risk Modeling via Direct Utility Maximization Using Numerical Quadrature
提出一种风险建模方法，通过非线性规划和数值积分直接求解期望效用最大化问题，明确指定联合概率密度函数，相比经验密度方法更精确。
Abstract An approach to risk modeling is developed which uses nonlinear programming and numerical integration to directly solve the expected utility maximization problem. The approach contrasts with earlier efforts in that, rather than using an empirical density function, a joint probability density function is explicitly specified. Comparisons are done showing that this approach yields more accurate solutions than the empirical density approach even when many points are sampled from the theoretical distribution.