CppAD: A C++ Algorithmic Differentiation Package
20171217

void CppAD::local::subgraph::subgraph_info::get_rev  (  const player< Base > *  play, 
const vector< size_t > &  dep_taddr,  
addr_t  i_dep,  
pod_vector< addr_t > &  subgraph  
) 
Get the subgraph corresponding to a dependent variables (and a selected set of independent variables).
Base  this operation sequence was recording using AD<Base>. 
play  is the operation sequence corresponding to the ADFun<Base> function. 
dep_taddr  is the vector mapping user dependent variable indices to the correpsonding variable in the recording. 
i_dep  is the user index for his dependent variable; that i_dep < n_dep_. 
subgraph  the input size and contents of this vector do not matter. Repeated calls with the same subgraph vector should reduce the amount of memory allocation. Upon return it contains the operator indices for the subgraph corresponding to the dependent and the selected independent variables. Only selected independent variable operators InvOp are included in the subgraph. Furthermore the operator indices in subgraph are unique; i.e., if i_op != j_op then subgraph[i_op] != subgraph[j_op]. 
Definition at line 84 of file get_rev.hpp.
Referenced by CppAD::local::subgraph::subgraph_sparsity().