Reliability Methods

This theory chapter explores local and global reliability methods in greater detail the overview in Uncertainty Quantification.

Local Reliability Methods

Local reliability methods include the Mean Value method and the family of most probable point (MPP) search methods. Each of these methods is gradient-based, employing local approximations and/or local optimization methods.

Mean Value

The Mean Value method (MV, also known as MVFOSM in [HM00]) is the simplest, least-expensive reliability method because it estimates the response means, response standard deviations, and all CDF/CCDF response-probability-reliability levels from a single evaluation of response functions and their gradients at the uncertain variable means. This approximation can have acceptable accuracy when the response functions are nearly linear and their distributions are approximately Gaussian, but can have poor accuracy in other situations.

The expressions for approximate response mean ${\mu }_g$ and approximate response variance ${\sigma }_g^2$ are

(54)$$\begin{array}{rlr}{\mu }_g& =g({\mu }_{\mathbf{x}})& \text{ (a)}\\ {\sigma }_g^2& =\sum _i\sum _{j}Cov(i,j)\frac{dg}{d{x}_i}({\mu }_{\mathbf{x}})\frac{dg}{d{x}_j}({\mu }_{\mathbf{x}})& \text{ (b)}\end{array}$$

where $\mathbf{x}$ are the uncertain values in the space of the original uncertain variables (“x-space”), $g(\mathbf{x})$ is the limit state function (the response function for which probability-response level pairs are needed), and the use of a linear Taylor series approximation is evident. These two moments are then used for mappings from response target to approximate reliability level ($\overline{z}\to \beta$) and from reliability target to approximate response level ($\overline{\beta }\to z$) using

(55)$$\begin{array}{rlr}\overline{z}\to \beta :& {\beta }_{\mathrm{C}\mathrm{D}\mathrm{F}}=\frac{{\mu }_g-\overline{z}}{{\sigma }_g},{\beta }_{\mathrm{C}\mathrm{C}\mathrm{D}\mathrm{F}}=\frac{\overline{z}-{\mu }_g}{{\sigma }_g}& \text{ (RIA)}\\ \overline{\beta }\to z:& z={\mu }_g-{\sigma }_g{\overline{\beta }}_{\mathrm{C}\mathrm{D}\mathrm{F}},z={\mu }_g+{\sigma }_g{\overline{\beta }}_{\mathrm{C}\mathrm{C}\mathrm{D}\mathrm{F}}& \text{ (PMA)}\end{array}$$

respectively, where ${\beta }_{\mathrm{C}\mathrm{D}\mathrm{F}}$ and ${\beta }_{\mathrm{C}\mathrm{C}\mathrm{D}\mathrm{F}}$ are the reliability indices corresponding to the cumulative and complementary cumulative distribution functions (CDF and CCDF), respectively.

With the introduction of second-order limit state information, MVSOSM calculates a second-order mean as

(56)$${\mu }_g=g({\mu }_{\mathbf{x}})+\frac{1}{2}\sum _i\sum _{j}Cov(i,j)\frac{d^{2}g}{d{x}_{i}d{x}_j}({\mu }_{\mathbf{x}})$$

This is commonly combined with a first-order variance ((b) in (54)), since second-order variance involves higher order distribution moments (skewness, kurtosis) [HM00] which are often unavailable.

The first-order CDF probability $p(g\le z)$, first-order CCDF probability $p(g>z)$, ${\beta }_{\mathrm{C}\mathrm{D}\mathrm{F}}$, and ${\beta }_{\mathrm{C}\mathrm{C}\mathrm{D}\mathrm{F}}$ are related to one another through

(57)$$\begin{array}{rlr}p(g\le z)& =\mathrm{\Phi }(-{\beta }_{\mathrm{C}\mathrm{D}\mathrm{F}})& \text{ (a)}\\ p(g>z)& =\mathrm{\Phi }(-{\beta }_{\mathrm{C}\mathrm{C}\mathrm{D}\mathrm{F}})& \text{ (b)}\\ {\beta }_{\mathrm{C}\mathrm{D}\mathrm{F}}& =-{\mathrm{\Phi }}^{-1}(p(g\le z))& \text{ (c)}\\ {\beta }_{\mathrm{C}\mathrm{C}\mathrm{D}\mathrm{F}}& =-{\mathrm{\Phi }}^{-1}(p(g>z))& \text{ (d)}\\ {\beta }_{\mathrm{C}\mathrm{D}\mathrm{F}}& =-{\beta }_{\mathrm{C}\mathrm{C}\mathrm{D}\mathrm{F}}& \text{ (e)}\\ p(g\le z)& =1-p(g>z)& \text{ (f)}\end{array}$$

where $\mathrm{\Phi }()$ is the standard normal cumulative distribution function, indicating the introduction of a Gaussian assumption on the output distributions. A common convention in the literature is to define $g$ in such a way that the CDF probability for a response level $z$ of zero (i.e., $p(g\le 0)$) is the response metric of interest. Dakota is not restricted to this convention and is designed to support CDF or CCDF mappings for general response, probability, and reliability level sequences.

With the Mean Value method, it is possible to obtain importance factors indicating the relative contribution of the input variables to the output variance. The importance factors can be viewed as an extension of linear sensitivity analysis combining deterministic gradient information with input uncertainty information, i.e. input variable standard deviations. The accuracy of the importance factors is contingent of the validity of the linear Taylor series approximation used to approximate the response quantities of interest. The importance factors are determined as follows for each of $n$ random variables:

$${\mathrm{I}\mathrm{m}\mathrm{p}\mathrm{o}\mathrm{r}\mathrm{t}\mathrm{F}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{o}\mathrm{r}}_i={[\frac{{\sigma }_{x_i}}{{\sigma }_g}\frac{dg}{d{x}_i}({\mu }_{\mathbf{x}})]}^2,i=1,\dots ,n$$

where it is evident that these importance factors correspond to the diagonal terms in (54), (b) normalized by the total response variance. In the case where the input variables are correlated resulting in off-diagonal terms for the input covariance, we can also compute a two-way importance factor as

$${\mathrm{I}\mathrm{m}\mathrm{p}\mathrm{o}\mathrm{r}\mathrm{t}\mathrm{F}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{o}\mathrm{r}}_{ij}=2\frac{{\sigma }_{x_{ij}}^2}{{\sigma }_g^2}\frac{dg}{d{x}_i}({\mu }_{\mathbf{x}})\frac{dg}{d{x}_j}({\mu }_{\mathbf{x}}),i=1,\dots ,n;j=1,\dots ,i-1$$

These two-way factors differ from the Sobol’ interaction terms that are computed in variance-based decomposition (refer to Global sensitivity analysis: variance-based decomposition) due to the non-orthogonality of the Taylor series basis. Due to this non-orthogonality, two-way importance factors may be negative, and due to normalization by the total response variance, the set of importance factors will always sum to one.

MPP Search Methods

All other local reliability methods solve an equality-constrained nonlinear optimization problem to compute a most probable point (MPP) and then integrate about this point to compute probabilities. The MPP search is performed in uncorrelated standard normal space (“u-space”) since it simplifies the probability integration: the distance of the MPP from the origin has the meaning of the number of input standard deviations separating the mean response from a particular response threshold. The transformation from correlated non-normal distributions (x-space) to uncorrelated standard normal distributions (u-space) is denoted as $\mathbf{u}=T(\mathbf{x})$ with the reverse transformation denoted as $\mathbf{x}=T^{-1}(\mathbf{u})$. These transformations are nonlinear in general, and possible approaches include the Rosenblatt [Ros52], Nataf [DKL86], and Box-Cox [BC64] transformations. The nonlinear transformations may also be linearized, and common approaches for this include the Rackwitz-Fiessler [RF78] two-parameter equivalent normal and the Chen-Lind [CL83] and Wu-Wirsching [WW87] three-parameter equivalent normals. Dakota employs the Nataf nonlinear transformation which is suitable for the common case when marginal distributions and a correlation matrix are provided, but full joint distributions are not known [1]. This transformation occurs in the following two steps. To transform between the original correlated x-space variables and correlated standard normals (“z-space”), a CDF matching condition is applied for each of the marginal distributions:

$$\mathrm{\Phi }(z_i)=F(x_i)$$

where $F()$ is the cumulative distribution function of the original probability distribution. Then, to transform between correlated z-space variables and uncorrelated u-space variables, the Cholesky factor $\mathbf{L}$ of a modified correlation matrix is used:

$$\mathbf{z}=\mathbf{L}\mathbf{u}$$

where the original correlation matrix for non-normals in x-space has been modified to represent the corresponding “warped” correlation in z-space [DKL86].

The forward reliability analysis algorithm of computing CDF/CCDF probability/reliability levels for specified response levels is called the reliability index approach (RIA), and the inverse reliability analysis algorithm of computing response levels for specified CDF/CCDF probability/reliability levels is called the performance measure approach (PMA) [TCP99]. The differences between the RIA and PMA formulations appear in the objective function and equality constraint formulations used in the MPP searches. For RIA, the MPP search for achieving the specified response level $\overline{z}$ is formulated as computing the minimum distance in u-space from the origin to the $\overline{z}$ contour of the limit state response function:

(58)$$\begin{array}{r}\begin{array}{rl}\mathtt{\text{minimize }}& {\mathbf{u}}^T\mathbf{u}\\ \mathtt{\text{subject to }}& G(\mathbf{u})=\overline{z}\end{array}\end{array}$$

where $\mathbf{u}$ is a vector centered at the origin in u-space and $g(\mathbf{x})\equiv G(\mathbf{u})$ by definition. For PMA, the MPP search for achieving the specified reliability level $\overline{\beta }$ or first-order probability level $\overline{p}$ is formulated as computing the minimum/maximum response function value corresponding to a prescribed distance from the origin in u-space:

(59)$$\begin{array}{r}\begin{array}{rl}\mathtt{\text{minimize }}& \pm G(\mathbf{u})\\ \mathtt{\text{subject to }}& {\mathbf{u}}^T\mathbf{u}={\overline{\beta }}^2\end{array}\end{array}$$

where $\overline{\beta }$ is computed from $\overline{p}$ using (c) and (d) in (57) in the latter case of a prescribed first-order probability level. For a specified generalized reliability level $\overline{{\beta }^{\ast }}$ or second-order probability level $\overline{p}$, the equality constraint is reformulated in terms of the generalized reliability index:

(60)$$\begin{array}{r}\begin{array}{rl}\mathtt{\text{minimize }}& \pm G(\mathbf{u})\\ \mathtt{\text{subject to }}& {\beta }^{\ast }(\mathbf{u})=\overline{{\beta }^{\ast }}\end{array}\end{array}$$

where $\overline{{\beta }^{\ast }}$ is computed from $\overline{p}$ using (271) (or its CCDF complement) in the latter case of a prescribed second-order probability level.

In the RIA case, the optimal MPP solution ${\mathbf{u}}^{\ast }$ defines the reliability index from $\beta =\pm \Vert {\mathbf{u}}^{\ast }{\Vert }_2$, which in turn defines the CDF/CCDF probabilities (using (a) and (b) in (57) in the case of first-order integration). The sign of $\beta$ is defined by

$$\begin{array}{r}G({\mathbf{u}}^{\ast })>G(\mathbf{0}):{\beta }_{\mathrm{C}\mathrm{D}\mathrm{F}}<0,{\beta }_{\mathrm{C}\mathrm{C}\mathrm{D}\mathrm{F}}>0\\ G({\mathbf{u}}^{\ast })<G(\mathbf{0}):{\beta }_{\mathrm{C}\mathrm{D}\mathrm{F}}>0,{\beta }_{\mathrm{C}\mathrm{C}\mathrm{D}\mathrm{F}}<0\end{array}$$

where $G(\mathbf{0})$ is the median limit state response computed at the origin in u-space [2] (where ${\beta }_{\mathrm{C}\mathrm{D}\mathrm{F}}$ = ${\beta }_{\mathrm{C}\mathrm{C}\mathrm{D}\mathrm{F}}$ = 0 and first-order $p(g\le z)$ = $p(g>z)$ = 0.5). In the PMA case, the sign applied to $G(\mathbf{u})$ (equivalent to minimizing or maximizing $G(\mathbf{u})$) is similarly defined by either $\overline{\beta }$ (for a specified reliability or first-order probability level) or from a $\overline{\beta }$ estimate [3] computed from $\overline{{\beta }^{\ast }}$ (for a specified generalized reliability or second-order probability level)

$$\begin{array}{r}{\overline{\beta }}_{\mathrm{C}\mathrm{D}\mathrm{F}}<0,{\overline{\beta }}_{\mathrm{C}\mathrm{C}\mathrm{D}\mathrm{F}}>0:\mathrm{m}\mathrm{a}\mathrm{x}\mathrm{i}\mathrm{m}\mathrm{i}\mathrm{z}\mathrm{e}G(\mathbf{u})\\ {\overline{\beta }}_{\mathrm{C}\mathrm{D}\mathrm{F}}>0,{\overline{\beta }}_{\mathrm{C}\mathrm{C}\mathrm{D}\mathrm{F}}<0:\mathrm{m}\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{m}\mathrm{i}\mathrm{z}\mathrm{e}G(\mathbf{u})\end{array}$$

where the limit state at the MPP ($G({\mathbf{u}}^{\ast })$) defines the desired response level result.

Limit state approximations

There are a variety of algorithmic variations that are available for use within RIA/PMA reliability analyses. First, one may select among several different limit state approximations that can be used to reduce computational expense during the MPP searches. Local, multipoint, and global approximations of the limit state are possible. [EAP+07] investigated local first-order limit state approximations, and [EB06] investigated local second-order and multipoint approximations. These techniques include:

1. a single Taylor series per response/reliability/probability level in x-space centered at the uncertain variable means. The first-order approach is commonly known as the Advanced Mean Value (AMV) method:

   (61)$$g(\mathbf{x})\cong g({\mu }_{\mathbf{x}})+{\mathrm{\nabla }}_{x}g({\mu }_{\mathbf{x}}{)}^T(\mathbf{x}-{\mu }_{\mathbf{x}})$$

   and the second-order approach has been named AMV${}^2$:

   (62)$$g(\mathbf{x})\cong g({\mu }_{\mathbf{x}})+{\mathrm{\nabla }}_{\mathbf{x}}g({\mu }_{\mathbf{x}}{)}^T(\mathbf{x}-{\mu }_{\mathbf{x}})+\frac{1}{2}(\mathbf{x}-{\mu }_{\mathbf{x}}{)}^T{\mathrm{\nabla }}_{\mathbf{x}}^{2}g({\mu }_{\mathbf{x}})(\mathbf{x}-{\mu }_{\mathbf{x}})$$

2. same as AMV/AMV${}^2$, except that the Taylor series is expanded in u-space. The first-order option has been termed the u-space AMV method:

   (63)$$G(\mathbf{u})\cong G({\mu }_{\mathbf{u}})+{\mathrm{\nabla }}_{u}G({\mu }_{\mathbf{u}}{)}^T(\mathbf{u}-{\mu }_{\mathbf{u}})$$

   where ${\mu }_{\mathbf{u}}=T({\mu }_{\mathbf{x}})$ and is nonzero in general, and the second-order option has been named the u-space AMV${}^2$ method:

   (64)$$G(\mathbf{u})\cong G({\mu }_{\mathbf{u}})+{\mathrm{\nabla }}_{\mathbf{u}}G({\mu }_{\mathbf{u}}{)}^T(\mathbf{u}-{\mu }_{\mathbf{u}})+\frac{1}{2}(\mathbf{u}-{\mu }_{\mathbf{u}}{)}^T{\mathrm{\nabla }}_{\mathbf{u}}^{2}G({\mu }_{\mathbf{u}})(\mathbf{u}-{\mu }_{\mathbf{u}})$$

3. an initial Taylor series approximation in x-space at the uncertain variable means, with iterative expansion updates at each MPP estimate (${\mathbf{x}}^{\ast }$) until the MPP converges. The first-order option is commonly known as AMV+:

   (65)$$g(\mathbf{x})\cong g({\mathbf{x}}^{\ast })+{\mathrm{\nabla }}_{x}g({\mathbf{x}}^{\ast }{)}^T(\mathbf{x}-{\mathbf{x}}^{\ast })$$

   and the second-order option has been named AMV${}^2$+:

   (66)$$g(\mathbf{x})\cong g({\mathbf{x}}^{\ast })+{\mathrm{\nabla }}_{\mathbf{x}}g({\mathbf{x}}^{\ast }{)}^T(\mathbf{x}-{\mathbf{x}}^{\ast })+\frac{1}{2}(\mathbf{x}-{\mathbf{x}}^{\ast }{)}^T{\mathrm{\nabla }}_{\mathbf{x}}^{2}g({\mathbf{x}}^{\ast })(\mathbf{x}-{\mathbf{x}}^{\ast })$$

4. same as AMV+/AMV${}^2$+, except that the expansions are performed in u-space. The first-order option has been termed the u-space AMV+ method.

   (67)$$G(\mathbf{u})\cong G({\mathbf{u}}^{\ast })+{\mathrm{\nabla }}_{u}G({\mathbf{u}}^{\ast }{)}^T(\mathbf{u}-{\mathbf{u}}^{\ast })$$

   and the second-order option has been named the u-space AMV${}^2$+ method:

   (68)$$G(\mathbf{u})\cong G({\mathbf{u}}^{\ast })+{\mathrm{\nabla }}_{\mathbf{u}}G({\mathbf{u}}^{\ast }{)}^T(\mathbf{u}-{\mathbf{u}}^{\ast })+\frac{1}{2}(\mathbf{u}-{\mathbf{u}}^{\ast }{)}^T{\mathrm{\nabla }}_{\mathbf{u}}^{2}G({\mathbf{u}}^{\ast })(\mathbf{u}-{\mathbf{u}}^{\ast })$$

5. a multipoint approximation in x-space. This approach involves a Taylor series approximation in intermediate variables where the powers used for the intermediate variables are selected to match information at the current and previous expansion points. Based on the two-point exponential approximation concept (TPEA, [FRB90]), the two-point adaptive nonlinearity approximation (TANA-3, [XG98]) approximates the limit state as:

   (69)$$g(\mathbf{x})\cong g({\mathbf{x}}_2)+\sum _{i=1}^n\frac{\partial g}{\partial x_i}({\mathbf{x}}_2)\frac{x_{i,2}^{1-p_i}}{p_i}(x_i^{p_i}-x_{i,2}^{p_i})+\frac{1}{2}ϵ(\mathbf{x})\sum _{i=1}^n(x_i^{p_i}-x_{i,2}^{p_i}{)}^2$$

   where $n$ is the number of uncertain variables and:

   (70)$$\begin{array}{r}\begin{array}{rlr}{p}_i& =1+\ln \left[\frac{\frac{\partial g}{\partial x_i}({\mathbf{x}}_1)}{\frac{\partial g}{\partial x_i}({\mathbf{x}}_2)}\right]/\ln \left[\frac{x_{i,1}}{x_{i,2}}\right]& \text{ (a)}\\ ϵ(\mathbf{x})& =\frac{H}{\sum _{i=1}^n(x_i^{p_i}-x_{i,1}^{p_i}{)}^2+\sum _{i=1}^n(x_i^{p_i}-x_{i,2}^{p_i}{)}^2}& \text{ (b)}\\ H& =2[g({\mathbf{x}}_1)-g({\mathbf{x}}_2)-\sum _{i=1}^n\frac{\partial g}{\partial x_i}({\mathbf{x}}_2)\frac{x_{i,2}^{1-p_i}}{p_i}(x_{i,1}^{p_i}-x_{i,2}^{p_i})]& \text{ (c)}\end{array}\end{array}$$

   and ${\mathbf{x}}_2$ and ${\mathbf{x}}_1$ are the current and previous MPP estimates in x-space, respectively. Prior to the availability of two MPP estimates, x-space AMV+ is used.

6. a multipoint approximation in u-space. The u-space TANA-3 approximates the limit state as:

   (71)$$G(\mathbf{u})\cong G({\mathbf{u}}_2)+\sum _{i=1}^n\frac{\partial G}{\partial u_i}({\mathbf{u}}_2)\frac{u_{i,2}^{1-p_i}}{p_i}(u_i^{p_i}-u_{i,2}^{p_i})+\frac{1}{2}ϵ(\mathbf{u})\sum _{i=1}^n(u_i^{p_i}-u_{i,2}^{p_i}{)}^2$$

   where:

   (72)$$\begin{array}{r}\begin{array}{rlr}{p}_i& =1+\ln \left[\frac{\frac{\partial G}{\partial u_i}({\mathbf{u}}_1)}{\frac{\partial G}{\partial u_i}({\mathbf{u}}_2)}\right]/\ln \left[\frac{u_{i,1}}{u_{i,2}}\right]& \text{ (a)}\\ ϵ(\mathbf{u})& =\frac{H}{\sum _{i=1}^n(u_i^{p_i}-u_{i,1}^{p_i}{)}^2+\sum _{i=1}^n(u_i^{p_i}-u_{i,2}^{p_i}{)}^2}& \text{ (b)}\\ H& =2[G({\mathbf{u}}_1)-G({\mathbf{u}}_2)-\sum _{i=1}^n\frac{\partial G}{\partial u_i}({\mathbf{u}}_2)\frac{u_{i,2}^{1-p_i}}{p_i}(u_{i,1}^{p_i}-u_{i,2}^{p_i})]& \text{ (c)}\end{array}\end{array}$$

   and ${\mathbf{u}}_2$ and ${\mathbf{u}}_1$ are the current and previous MPP estimates in u-space, respectively. Prior to the availability of two MPP estimates, u-space AMV+ is used.

7. the MPP search on the original response functions without the use of any approximations. Combining this option with first-order and second-order integration approaches (see next section) results in the traditional first-order and second-order reliability methods (FORM and SORM).

The Hessian matrices in AMV${}^2$ and AMV${}^2$+ may be available analytically, estimated numerically, or approximated through quasi-Newton updates. The selection between x-space or u-space for performing approximations depends on where the approximation will be more accurate, since this will result in more accurate MPP estimates (AMV, AMV${}^2$) or faster convergence (AMV+, AMV${}^2$+, TANA). Since this relative accuracy depends on the forms of the limit state $g(x)$ and the transformation $T(x)$ and is therefore application dependent in general, Dakota supports both options. A concern with approximation-based iterative search methods (i.e., AMV+, AMV${}^2$+ and TANA) is the robustness of their convergence to the MPP. It is possible for the MPP iterates to oscillate or even diverge. However, to date, this occurrence has been relatively rare, and Dakota contains checks that monitor for this behavior. Another concern with TANA is numerical safeguarding (e.g., the possibility of raising

negative $x_i$ or $u_i$ values to nonintegral $p_i$ exponents in (69), (70) (b) – (71), and (72) (b) and (c).

Safeguarding involves offseting negative $x_i$ or $u_i$ and, for potential numerical difficulties with the logarithm ratios in (b) in each of (70) and (72), reverting to either the linear ($p_i=1$) or reciprocal ($p_i=-1$) approximation based on which approximation has lower error in $\frac{\partial g}{\partial x_i}({\mathbf{x}}_1)$ or $\frac{\partial G}{\partial u_i}({\mathbf{u}}_1)$.

Probability integrations

The second algorithmic variation involves the integration approach for computing probabilities at the MPP, which can be selected to be first-order ((a) and (b) in (57)) or second-order integration. Second-order integration involves applying a curvature correction [Bre84, HR88, Hon99]. Breitung applies a correction based on asymptotic analysis [Bre84]:

(73)$$p=\mathrm{\Phi }(-{\beta }_p)\prod _{i=1}^{n-1}\frac{1}{\sqrt{1+{\beta }_p{\kappa }_i}}$$

where ${\kappa }_i$ are the principal curvatures of the limit state function (the eigenvalues of an orthonormal transformation of ${\mathrm{\nabla }}_{\mathbf{u}}^{2}G$, taken positive for a convex limit state) and ${\beta }_p\ge 0$ (a CDF or CCDF probability correction is selected to obtain the correct sign for ${\beta }_p$). An alternate correction in [HR88] is consistent in the asymptotic regime (${\beta }_p\to \mathrm{\infty }$) but does not collapse to first-order integration for ${\beta }_p=0$:

(74)$$p=\mathrm{\Phi }(-{\beta }_p)\prod _{i=1}^{n-1}\frac{1}{\sqrt{1+\psi (-{\beta }_p){\kappa }_i}}$$

where $\psi ()=\frac{\varphi ()}{\mathrm{\Phi }()}$ and $\varphi ()$ is the standard normal density function. [Hon99] applies further corrections to (74) based on point concentration methods. At this time, all three approaches are available within the code, but the Hohenbichler-Rackwitz correction is used by default (switching the correction is a compile-time option in the source code and has not been exposed in the input specification).

Hessian approximations

To use a second-order Taylor series or a second-order integration when second-order information (${\mathrm{\nabla }}_{\mathbf{x}}^{2}g$, ${\mathrm{\nabla }}_{\mathbf{u}}^{2}G$, and/or $\kappa$) is not directly available, one can estimate the missing information using finite differences or approximate it through use of quasi-Newton approximations. These procedures will often be needed to make second-order approaches practical for engineering applications.

In the finite difference case, numerical Hessians are commonly computed using either first-order forward differences of gradients using

$${\mathrm{\nabla }}^{2}g(\mathbf{x})\cong \frac{\mathrm{\nabla }g(\mathbf{x}+h{\mathbf{e}}_i)-\mathrm{\nabla }g(\mathbf{x})}{h}$$

to estimate the $i^{th}$ Hessian column when gradients are analytically available, or second-order differences of function values using

$${\mathrm{\nabla }}^{2}g(\mathbf{x})\cong \frac{g(\mathbf{x}+h{\mathbf{e}}_i+h{\mathbf{e}}_j)-g(\mathbf{x}+h{\mathbf{e}}_i-h{\mathbf{e}}_j)-g(\mathbf{x}-h{\mathbf{e}}_i+h{\mathbf{e}}_j)+g(\mathbf{x}-h{\mathbf{e}}_i-h{\mathbf{e}}_j)}{4h^2}$$

to estimate the $i{j}^{th}$ Hessian term when gradients are not directly available. This approach has the advantage of locally-accurate Hessians for each point of interest (which can lead to quadratic convergence rates in discrete Newton methods), but has the disadvantage that numerically estimating each of the matrix terms can be expensive.

Quasi-Newton approximations, on the other hand, do not reevaluate all of the second-order information for every point of interest. Rather, they accumulate approximate curvature information over time using secant updates. Since they utilize the existing gradient evaluations, they do not require any additional function evaluations for evaluating the Hessian terms. The quasi-Newton approximations of interest include the Broyden-Fletcher-Goldfarb-Shanno (BFGS) update

(75)$${\mathbf{B}}_{k+1}={\mathbf{B}}_k-\frac{{\mathbf{B}}_k{\mathbf{s}}_k{\mathbf{s}}_k^T{\mathbf{B}}_k}{{\mathbf{s}}_k^T{\mathbf{B}}_k{\mathbf{s}}_k}+\frac{{\mathbf{y}}_k{\mathbf{y}}_k^T}{{\mathbf{y}}_k^T{\mathbf{s}}_k}$$

which yields a sequence of symmetric positive definite Hessian approximations, and the Symmetric Rank 1 (SR1) update

$${\mathbf{B}}_{k+1}={\mathbf{B}}_k+\frac{({\mathbf{y}}_k-{\mathbf{B}}_k{\mathbf{s}}_k)({\mathbf{y}}_k-{\mathbf{B}}_k{\mathbf{s}}_k{)}^T}{({\mathbf{y}}_k-{\mathbf{B}}_k{\mathbf{s}}_k{)}^T{\mathbf{s}}_k}$$

which yields a sequence of symmetric, potentially indefinite, Hessian approximations. ${\mathbf{B}}_k$ is the $k^{th}$ approximation to the Hessian ${\mathrm{\nabla }}^{2}g$, ${\mathbf{s}}_k={\mathbf{x}}_{k+1}-{\mathbf{x}}_k$ is the step and ${\mathbf{y}}_k=\mathrm{\nabla }{g}_{k+1}-\mathrm{\nabla }{g}_k$ is the corresponding yield in the gradients. The selection of BFGS versus SR1 involves the importance of retaining positive definiteness in the Hessian approximations; if the procedure does not require it, then the SR1 update can be more accurate if the true Hessian is not positive definite. Initial scalings for ${\mathbf{B}}_0$ and numerical safeguarding techniques (damped BFGS, update skipping) are described in [EB06].

Optimization algorithms

The next algorithmic variation involves the optimization algorithm selection for solving Eqs. (58) and (59). The Hasofer-Lind Rackwitz-Fissler (HL-RF) algorithm [HM00] is a classical approach that has been broadly applied. It is a Newton-based approach lacking line search/trust region globalization, and is generally regarded as computationally efficient but occasionally unreliable. Dakota takes the approach of employing robust, general-purpose optimization algorithms with provable convergence properties. In particular, we employ the sequential quadratic programming (SQP) and nonlinear interior-point (NIP) optimization algorithms from the NPSOL [GMSW86] and OPT++ [MOHW07] libraries, respectively.

Warm Starting of MPP Searches

The final algorithmic variation for local reliability methods involves the use of warm starting approaches for improving computational efficiency. [EAP+07] describes the acceleration of MPP searches through warm starting with approximate iteration increment, with $z/p/\beta$ level increment, and with design variable increment. Warm started data includes the expansion point and associated response values and the MPP optimizer initial guess. Projections are used when an increment in $z/p/\beta$ level or design variables occurs. Warm starts were consistently effective in [EAP+07], with greater effectiveness for smaller parameter changes, and are used by default in Dakota.

Global Reliability Methods

Local reliability methods, while computationally efficient, have well-known failure mechanisms. When confronted with a limit state function that is nonsmooth, local gradient-based optimizers may stall due to gradient inaccuracy and fail to converge to an MPP. Moreover, if the limit state is multimodal (multiple MPPs), then a gradient-based local method can, at best, locate only one local MPP solution. Finally, a linear (Eqs. (a) and (b) in (57)) or parabolic (Eqs. (73) – (74)) approximation to the limit state at this MPP may fail to adequately capture the contour of a highly nonlinear limit state.

A reliability analysis method that is both efficient when applied to expensive response functions and accurate for a response function of any arbitrary shape is needed. This section develops such a method based on efficient global optimization [JSW98] (EGO) to the search for multiple points on or near the limit state throughout the random variable space. By locating multiple points on the limit state, more complex limit states can be accurately modeled, resulting in a more accurate assessment of the reliability. It should be emphasized here that these multiple points exist on a single limit state. Because of its roots in efficient global optimization, this method of reliability analysis is called efficient global reliability analysis (EGRA) [BES+07]. The following two subsections describe two capabilities that are incorporated into the EGRA algorithm: importance sampling and EGO.

Importance Sampling

An alternative to MPP search methods is to directly perform the probability integration numerically by sampling the response function. Sampling methods do not rely on a simplifying approximation to the shape of the limit state, so they can be more accurate than FORM and SORM, but they can also be prohibitively expensive because they generally require a large number of response function evaluations. Importance sampling methods reduce this expense by focusing the samples in the important regions of the uncertain space. They do this by centering the sampling density function at the MPP rather than at the mean. This ensures the samples will lie the region of interest, thus increasing the efficiency of the sampling method. Adaptive importance sampling (AIS) further improves the efficiency by adaptively updating the sampling density function. Multimodal adaptive importance sampling [DM98, ZMMM02] is a variation of AIS that allows for the use of multiple sampling densities making it better suited for cases where multiple sections of the limit state are highly probable.

Note that importance sampling methods require that the location of at least one MPP be known because it is used to center the initial sampling density. However, current gradient-based, local search methods used in MPP search may fail to converge or may converge to poor solutions for highly nonlinear problems, possibly making these methods inapplicable. As the next section describes, EGO is a global optimization method that does not depend on the availability of accurate gradient information, making convergence more reliable for nonsmooth response functions. Moreover, EGO has the ability to locate multiple failure points, which would provide multiple starting points and thus a good multimodal sampling density for the initial steps of multimodal AIS. The resulting Gaussian process model is accurate in the vicinity of the limit state, thereby providing an inexpensive surrogate that can be used to provide response function samples. As will be seen, using EGO to locate multiple points along the limit state, and then using the resulting Gaussian process model to provide function evaluations in multimodal AIS for the probability integration, results in an accurate and efficient reliability analysis tool.

Efficient Global Optimization

Note

Chapter Efficient Global Optimization has been substantially revised to discuss EGO/Bayesian optimization theory.

Expected Feasibility Function

The expected improvement function provides an indication of how much the true value of the response at a point can be expected to be less than the current best solution. It therefore makes little sense to apply this to the forward reliability problem where the goal is not to minimize the response, but rather to find where it is equal to a specified threshold value. The expected feasibility function (EFF) is introduced here to provide an indication of how well the true value of the response is expected to satisfy the equality constraint $G(\mathbf{u})=\overline{z}$. Inspired by the contour estimation work in [RDG08], this expectation can be calculated in a similar fashion as (262) by integrating over a region in the immediate vicinity of the threshold value $\overline{z}\pm ϵ$:

$$EF(\hat{G}(\mathbf{u}))={\int }_{z-ϵ}^{z+ϵ}[ϵ-|\overline{z}-G|]\hat{G}(\mathbf{u})dG$$

where $G$ denotes a realization of the distribution $\hat{G}$, as before. Allowing $z^{+}$ and $z^{-}$ to denote $\overline{z}\pm ϵ$, respectively, this integral can be expressed analytically as:

(76)$$\begin{array}{r}\begin{array}{rl}EF(\hat{G}(\mathbf{u}))& =({\mu }_G-\overline{z})[2\mathrm{\Phi }\left(\frac{\overline{z}-{\mu }_G}{{\sigma }_G}\right)-\mathrm{\Phi }\left(\frac{z^{-}-{\mu }_G}{{\sigma }_G}\right)-\mathrm{\Phi }\left(\frac{z^{+}-{\mu }_G}{{\sigma }_G}\right)]\\ & -{\sigma }_G[2\varphi \left(\frac{\overline{z}-{\mu }_G}{{\sigma }_G}\right)-\varphi \left(\frac{z^{-}-{\mu }_G}{{\sigma }_G}\right)-\varphi \left(\frac{z^{+}-{\mu }_G}{{\sigma }_G}\right)]\\ & +ϵ[\mathrm{\Phi }\left(\frac{z^{+}-{\mu }_G}{{\sigma }_G}\right)-\mathrm{\Phi }\left(\frac{z^{-}-{\mu }_G}{{\sigma }_G}\right)]\end{array}\end{array}$$

where $ϵ$ is proportional to the standard deviation of the GP predictor ($ϵ\propto {\sigma }_G$). In this case, $z^{-}$, $z^{+}$, ${\mu }_G$, ${\sigma }_G$, and $ϵ$ are all functions of the location $\mathbf{u}$, while $\overline{z}$ is a constant. Note that the EFF provides the same balance between exploration and exploitation as is captured in the EIF. Points where the expected value is close to the threshold (${\mu }_G\approx \overline{z}$) and points with a large uncertainty in the prediction will have large expected feasibility values.

[1]

If joint distributions are known, then the Rosenblatt transformation is preferred.

[2]

It is not necessary to explicitly compute the median response since the sign of the inner product $⟨{\mathbf{u}}^{\ast },{\mathrm{\nabla }}_{\mathbf{u}}G⟩$ can be used to determine the orientation of the optimal response with respect to the median response.

[3]

computed by inverting the second-order probability relationships described in Probability integrations at the current ${\mathbf{u}}^{\ast }$ iterate.