\(\newcommand{\W}[1]{ \; #1 \; }\) \(\newcommand{\R}[1]{ {\rm #1} }\) \(\newcommand{\B}[1]{ {\bf #1} }\) \(\newcommand{\D}[2]{ \frac{\partial #1}{\partial #2} }\) \(\newcommand{\DD}[3]{ \frac{\partial^2 #1}{\partial #2 \partial #3} }\) \(\newcommand{\Dpow}[2]{ \frac{\partial^{#1}}{\partial {#2}^{#1}} }\) \(\newcommand{\dpow}[2]{ \frac{ {\rm d}^{#1}}{{\rm d}\, {#2}^{#1}} }\)

subgraph_jac_rev

View page source

Compute Sparse Jacobians Using Subgraphs

Syntax

f . subgraph_jac_rev ( x , subset )

f . subgraph_jac_rev (

      select_domain , select_range , x , matrix_out

)

See Also

clear_subgraph .

Purpose

We use \(F : \B{R}^n \rightarrow \B{R}^m\) to denote the function corresponding to f . Here n is the Domain size, and m is the Range size, or f . The syntax above takes advantage of sparsity when computing the Jacobian

\[J(x) = F^{(1)} (x)\]

The first syntax requires one to know what which elements of the Jacobian they want to compute. The second syntax computes the sparsity pattern and the value of the Jacobian at the same time. If one only wants the sparsity pattern, it should be faster to use subgraph_sparsity .

Method

This routine uses a subgraph technique. To be specific, for each dependent variable, it creates a subgraph of the operation sequence containing the variables that affect the dependent variable. This avoids the overhead of performing set operations that is inherent in other methods for computing sparsity patterns.

BaseVector

The type BaseVector is a SimpleVector class with elements of type Base .

SizeVector

The type SizeVector is a SimpleVector class with elements of type size_t .

BoolVector

The type BoolVector is a SimpleVector class with elements of type bool .

f

This object has prototype

ADFun < Base > f

Note that the Taylor coefficients stored in f are affected by this operation; see Uses Forward below.

x

This argument has prototype

const BaseVector & x

It is the value of x at which we are computing the Jacobian.

Uses Forward

After each call to Forward , the object f contains the corresponding Taylor coefficients . After a call to sparse_jac_forward or sparse_jac_rev , the zero order coefficients correspond to

f . Forward (0, x )

All the other forward mode coefficients are unspecified.

subset

This argument has prototype

sparse_rcv < SizeVector , BaseVector >& subset

Its row size is subset . nr () == m , and its column size is subset . nc () == n . It specifies which elements of the Jacobian are computed. The input elements in its value vector subset . val () do not matter. Upon return it contains the value of the corresponding elements of the Jacobian.

select_domain

The argument select_domain has prototype

const BoolVector & select_domain

It has size \(n\) and specifies which independent variables to include.

select_range

The argument select_range has prototype

const BoolVector & select_range

It has size \(m\) and specifies which components of the range to include in the calculation. A subgraph is built for each dependent variable and the selected set of independent variables.

matrix_out

This argument has prototype

sparse_rcv < SizeVector , BaseVector >& matrix_out

This input value of matrix_out does not matter. Upon return matrix_out is sparse matrix representation of \(F^{(1)} (x)\). The matrix has \(m\) rows, \(n\) columns. If select_domain [ j ] is true, select_range [ i ] is true, and \(F_i (x)\) depends on \(x_j\), then the pair \((i, j)\) is in matrix_out . For each k = 0 , …, matrix_out . nnz () , let

      i = matrix_out . row ()[ k ]

      j = matrix_out . col ()[ k ]

      v = matrix_out . val ()[ k ]

It follows that the partial of \(F_i (x)\) with respect to \(x_j\) is equal to \(v\).

Example

The files subgraph_jac_rev.cpp and subgraph_hes2jac.cpp are examples and tests using subgraph_jac_rev . They returns true for success and false for failure.