First-order second-moment method

In probability theory, the first-order second-moment (FOSM) method, also referenced as mean value first-order second-moment (MVFOSM) method, is a probabilistic method to determine the stochastic moments of a function with random input variables. The name is based on the derivation, which uses a first-order Taylor series and the first and second moments of the input variables.^[1]

Consider the objective function ${\displaystyle g(x)}${\displaystyle g(x)}, where the input vector ${\displaystyle x}${\displaystyle x} is a realization of the random vector ${\displaystyle X}${\displaystyle X} with probability density function ${\displaystyle f_{X}(x)}${\displaystyle f_{X}(x)}. Because ${\displaystyle X}${\displaystyle X} is randomly distributed, ${\displaystyle g}${\displaystyle g} is also randomly distributed. Following the FOSM method, the mean value of ${\displaystyle g}${\displaystyle g} is approximated by

${\displaystyle \mu _{g}\approx g(\mu )}${\displaystyle \mu _{g}\approx g(\mu )}

The variance of ${\displaystyle g}${\displaystyle g} is approximated by

${\displaystyle \sigma _{g}^{2}\approx \sum _{i=1}^{n}\sum _{j=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}{\frac {\partial g(\mu )}{\partial x_{j}}}\operatorname {cov} \left(X_{i},X_{j}\right)}${\displaystyle \sigma _{g}^{2}\approx \sum _{i=1}^{n}\sum _{j=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}{\frac {\partial g(\mu )}{\partial x_{j}}}\operatorname {cov} \left(X_{i},X_{j}\right)}

where ${\displaystyle n}${\displaystyle n} is the length/dimension of ${\displaystyle x}${\displaystyle x} and ${\textstyle {\frac {\partial g(\mu )}{\partial x_{i}}}}${\textstyle {\frac {\partial g(\mu )}{\partial x_{i}}}} is the partial derivative of ${\displaystyle g}${\displaystyle g} at the mean vector ${\displaystyle \mu }${\displaystyle \mu } with respect to the i-th entry of ${\displaystyle x}${\displaystyle x}. More accurate, second-order second-moment approximations are also available ^[2]

The objective function is approximated by a Taylor series at the mean vector ${\displaystyle \mu }${\displaystyle \mu }.

${\displaystyle g(x)=g(\mu )+\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})+{\frac {1}{2}}\sum _{i=1}^{n}\sum _{j=1}^{n}{\frac {\partial ^{2}g(\mu )}{\partial x_{i},円\partial x_{j}}}(x_{i}-\mu _{i})(x_{j}-\mu _{j})+\cdots }${\displaystyle g(x)=g(\mu )+\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})+{\frac {1}{2}}\sum _{i=1}^{n}\sum _{j=1}^{n}{\frac {\partial ^{2}g(\mu )}{\partial x_{i},円\partial x_{j}}}(x_{i}-\mu _{i})(x_{j}-\mu _{j})+\cdots }

The mean value of ${\displaystyle g}${\displaystyle g} is given by the integral

${\displaystyle \mu _{g}=E[g(x)]=\int _{-\infty }^{\infty }g(x)f_{X}(x),円dx}${\displaystyle \mu _{g}=E[g(x)]=\int _{-\infty }^{\infty }g(x)f_{X}(x),円dx}

Inserting the first-order Taylor series yields

${\displaystyle {\begin{aligned}\mu _{g}&\approx \int _{-\infty }^{\infty }\left[g(\mu )+\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})\right]f_{X}(x),円dx\\&=\int _{-\infty }^{\infty }g(\mu )f_{X}(x),円dx+\int _{-\infty }^{\infty }\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})f_{X}(x),円dx\\&=g(\mu )\underbrace {\int _{-\infty }^{\infty }f_{X}(x),円dx} _{1}+\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}\underbrace {\int _{-\infty }^{\infty }(x_{i}-\mu _{i})f_{X}(x),円dx} _{0}\\&=g(\mu ).\end{aligned}}}${\displaystyle {\begin{aligned}\mu _{g}&\approx \int _{-\infty }^{\infty }\left[g(\mu )+\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})\right]f_{X}(x),円dx\\&=\int _{-\infty }^{\infty }g(\mu )f_{X}(x),円dx+\int _{-\infty }^{\infty }\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})f_{X}(x),円dx\\&=g(\mu )\underbrace {\int _{-\infty }^{\infty }f_{X}(x),円dx} _{1}+\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}\underbrace {\int _{-\infty }^{\infty }(x_{i}-\mu _{i})f_{X}(x),円dx} _{0}\\&=g(\mu ).\end{aligned}}}

The variance of ${\displaystyle g}${\displaystyle g} is given by the integral

${\displaystyle \sigma _{g}^{2}=E\left([g(x)-\mu _{g}]^{2}\right)=\int _{-\infty }^{\infty }[g(x)-\mu _{g}]^{2}f_{X}(x),円dx.}${\displaystyle \sigma _{g}^{2}=E\left([g(x)-\mu _{g}]^{2}\right)=\int _{-\infty }^{\infty }[g(x)-\mu _{g}]^{2}f_{X}(x),円dx.}

According to the computational formula for the variance, this can be written as

${\displaystyle \sigma _{g}^{2}=E\left([g(x)-\mu _{g}]^{2}\right)=E\left(g(x)^{2}\right)-\mu _{g}^{2}=\int _{-\infty }^{\infty }g(x)^{2}f_{X}(x),円dx-\mu _{g}^{2}}${\displaystyle \sigma _{g}^{2}=E\left([g(x)-\mu _{g}]^{2}\right)=E\left(g(x)^{2}\right)-\mu _{g}^{2}=\int _{-\infty }^{\infty }g(x)^{2}f_{X}(x),円dx-\mu _{g}^{2}}

Inserting the Taylor series yields

${\displaystyle {\begin{aligned}\sigma _{g}^{2}&\approx \int _{-\infty }^{\infty }\left[g(\mu )+\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})\right]^{2}f_{X}(x),円dx-\mu _{g}^{2}\\&=\int _{-\infty }^{\infty }\left\{g(\mu )^{2}+2g_{\mu }\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})+\left[\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})\right]^{2}\right\}f_{X}(x),円dx-\mu _{g}^{2}\\&=\int _{-\infty }^{\infty }g(\mu )^{2}f_{X}(x),円dx+\int _{-\infty }^{\infty }2,円g_{\mu }\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})f_{X}(x),円dx\\&\quad {}+\int _{-\infty }^{\infty }\left[\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})\right]^{2}f_{X}(x),円dx-\mu _{g}^{2}\\&=g_{\mu }^{2}\underbrace {\int _{-\infty }^{\infty }f_{X}(x),円dx} _{1}+2g_{\mu }\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}\underbrace {\int _{-\infty }^{\infty }(x_{i}-\mu _{i})f_{X}(x),円dx} _{0}\\&\quad {}+\int _{-\infty }^{\infty }\left[\sum _{i=1}^{n}\sum _{j=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}{\frac {\partial g(\mu )}{\partial x_{j}}}(x_{i}-\mu _{i})(x_{j}-\mu _{j})\right]f_{X}(x),円dx-\mu _{g}^{2}\\&=\underbrace {g(\mu )^{2}} _{\mu _{g}^{2}}+\sum _{i=1}^{n}\sum _{j=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}{\frac {\partial g(\mu )}{\partial x_{j}}}\underbrace {\int _{-\infty }^{\infty }(x_{i}-\mu _{i})(x_{j}-\mu _{j})f_{X}(x),円dx} _{\operatorname {cov} \left(X_{i},X_{j}\right)}-\mu _{g}^{2}\\&=\sum _{i=1}^{n}\sum _{j=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}{\frac {\partial g(\mu )}{\partial x_{j}}}\operatorname {cov} \left(X_{i},X_{j}\right).\end{aligned}}}${\displaystyle {\begin{aligned}\sigma _{g}^{2}&\approx \int _{-\infty }^{\infty }\left[g(\mu )+\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})\right]^{2}f_{X}(x),円dx-\mu _{g}^{2}\\&=\int _{-\infty }^{\infty }\left\{g(\mu )^{2}+2g_{\mu }\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})+\left[\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})\right]^{2}\right\}f_{X}(x),円dx-\mu _{g}^{2}\\&=\int _{-\infty }^{\infty }g(\mu )^{2}f_{X}(x),円dx+\int _{-\infty }^{\infty }2,円g_{\mu }\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})f_{X}(x),円dx\\&\quad {}+\int _{-\infty }^{\infty }\left[\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}(x_{i}-\mu _{i})\right]^{2}f_{X}(x),円dx-\mu _{g}^{2}\\&=g_{\mu }^{2}\underbrace {\int _{-\infty }^{\infty }f_{X}(x),円dx} _{1}+2g_{\mu }\sum _{i=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}\underbrace {\int _{-\infty }^{\infty }(x_{i}-\mu _{i})f_{X}(x),円dx} _{0}\\&\quad {}+\int _{-\infty }^{\infty }\left[\sum _{i=1}^{n}\sum _{j=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}{\frac {\partial g(\mu )}{\partial x_{j}}}(x_{i}-\mu _{i})(x_{j}-\mu _{j})\right]f_{X}(x),円dx-\mu _{g}^{2}\\&=\underbrace {g(\mu )^{2}} _{\mu _{g}^{2}}+\sum _{i=1}^{n}\sum _{j=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}{\frac {\partial g(\mu )}{\partial x_{j}}}\underbrace {\int _{-\infty }^{\infty }(x_{i}-\mu _{i})(x_{j}-\mu _{j})f_{X}(x),円dx} _{\operatorname {cov} \left(X_{i},X_{j}\right)}-\mu _{g}^{2}\\&=\sum _{i=1}^{n}\sum _{j=1}^{n}{\frac {\partial g(\mu )}{\partial x_{i}}}{\frac {\partial g(\mu )}{\partial x_{j}}}\operatorname {cov} \left(X_{i},X_{j}\right).\end{aligned}}}

The following abbreviations are introduced.

${\displaystyle {\begin{aligned}g_{\mu }&=g(\mu ),&g_{,i}&={\frac {\partial g(\mu )}{\partial x_{i}}},&g_{,ij}&={\frac {\partial ^{2}g(\mu )}{\partial x_{i},円\partial x_{j}}},&\mu _{i,j}&=E\left[(x_{i}-\mu _{i})^{j}\right]\end{aligned}}}${\displaystyle {\begin{aligned}g_{\mu }&=g(\mu ),&g_{,i}&={\frac {\partial g(\mu )}{\partial x_{i}}},&g_{,ij}&={\frac {\partial ^{2}g(\mu )}{\partial x_{i},円\partial x_{j}}},&\mu _{i,j}&=E\left[(x_{i}-\mu _{i})^{j}\right]\end{aligned}}}

In the following, the entries of the random vector ${\displaystyle X}${\displaystyle X} are assumed to be independent. Considering also the second-order terms of the Taylor expansion, the approximation of the mean value is given by

${\displaystyle \mu _{g}\approx g_{\mu }+{\frac {1}{2}}\sum _{i=1}^{n}g_{,ii}\;\mu _{i,2}}${\displaystyle \mu _{g}\approx g_{\mu }+{\frac {1}{2}}\sum _{i=1}^{n}g_{,ii}\;\mu _{i,2}}

The incomplete second-order approximation (ISOA^[3] ) of the variance is given by

${\displaystyle {\begin{aligned}\sigma _{g}^{2}\approx {}g_{\mu }^{2}&+\sum _{i=1}^{n}g_{,i}^{2},円\mu _{i,2}+{\frac {1}{4}}\sum _{i=1}^{n}g_{,ii}^{2},円\mu _{i,4}+g_{\mu }\sum _{i=1}^{n}g_{,ii},円\mu _{i,2}+\sum _{i=1}^{n}g_{,i},円g_{,ii},円\mu _{i,3}\\&+{\frac {1}{2}}\sum _{i=1}^{n}\sum _{j=i+1}^{n}g_{,ii},円g_{,jj},円\mu _{i,2},円\mu _{j,2}+\sum _{i=1}^{n}\sum _{j=i+1}^{n}g_{,ij}^{2},円\mu _{i,2},円\mu _{j,2}-\mu _{g}^{2}\end{aligned}}}${\displaystyle {\begin{aligned}\sigma _{g}^{2}\approx {}g_{\mu }^{2}&+\sum _{i=1}^{n}g_{,i}^{2},円\mu _{i,2}+{\frac {1}{4}}\sum _{i=1}^{n}g_{,ii}^{2},円\mu _{i,4}+g_{\mu }\sum _{i=1}^{n}g_{,ii},円\mu _{i,2}+\sum _{i=1}^{n}g_{,i},円g_{,ii},円\mu _{i,3}\\&+{\frac {1}{2}}\sum _{i=1}^{n}\sum _{j=i+1}^{n}g_{,ii},円g_{,jj},円\mu _{i,2},円\mu _{j,2}+\sum _{i=1}^{n}\sum _{j=i+1}^{n}g_{,ij}^{2},円\mu _{i,2},円\mu _{j,2}-\mu _{g}^{2}\end{aligned}}}

The skewness of ${\displaystyle g}${\displaystyle g} can be determined from the third central moment ${\displaystyle \mu _{g,3}}${\displaystyle \mu _{g,3}}. When considering only linear terms of the Taylor series, but higher-order moments, the third central moment is approximated by

${\displaystyle \mu _{g,3}\approx \sum _{i=1}^{n}g_{,i}^{3}\;\mu _{i,3}}${\displaystyle \mu _{g,3}\approx \sum _{i=1}^{n}g_{,i}^{3}\;\mu _{i,3}}

For the second-order approximations of the third central moment as well as for the derivation of all higher-order approximations see Appendix D of Ref.^[3] Taking into account the quadratic terms of the Taylor series and the third moments of the input variables is referred to as second-order third-moment method.^[4] However, the full second-order approach of the variance (given above) also includes fourth-order moments of input parameters,^[5] the full second-order approach of the skewness 6th-order moments,^{[3] [6]} and the full second-order approach of the kurtosis up to 8th-order moments.^[6]

There are several examples in the literature where the FOSM method is employed to estimate the stochastic distribution of the buckling load of axially compressed structures (see e.g. Ref.^{[7] [8] [9] [10]} ). For structures which are very sensitive to deviations from the ideal structure (like cylindrical shells) it has been proposed to use the FOSM method as a design approach. Often the applicability is checked by comparison with a Monte Carlo simulation. Two comprehensive application examples of the full second-order method specifically oriented towards the fatigue crack growth in a metal railway axle are discussed and checked by comparison with a Monte Carlo simulation in Ref.^{[5] [6]}

In engineering practice, the objective function often is not given as analytic expression, but for instance as a result of a finite-element simulation. Then the derivatives of the objective function need to be estimated by the central differences method. The number of evaluations of the objective function equals ${\displaystyle 2n+1}${\displaystyle 2n+1}. Depending on the number of random variables this still can mean a significantly smaller number of evaluations than performing a Monte Carlo simulation. However, when using the FOSM method as a design procedure, a lower bound shall be estimated, which is actually not given by the FOSM approach. Therefore, a type of distribution needs to be assumed for the distribution of the objective function, taking into account the approximated mean value and standard deviation.

• Probabilistic models
• Randomized algorithms