在C ++ API中使用Octave内置函数mldivide

Question

I am programming in C++ and want to use the powerful built-in functions in Octave. I followed the standalone program guideline on Octave website. I can run the function norm (which is called as Fnorm in C++) with the sample code successfully. Now I want to use the function mldivide to solve a linear equation.

#include <iostream>
#include <octave/oct.h>
#include <octave/builtin-defun-decls.h>

octave_value_list input;
octave_value_list retval;

int main (void) {

Matrix A(4,4);

for (octave_idx_type i = 0; i < 4; i++)
   for (octave_idx_type j = 0; j < 4; j++)
    A(i,j) = 1.0 / (static_cast<double> (i) +static_cast<double> ( j ) + 1.0 ) ;

    ColumnVector b(4,1.0);
    input.append(A);
    input.append(b);

    retval=Fmldivide(input);
    ColumnVector x =retval(0).column_vector_value();

    std::cout << "A = " << std::endl << A << std::endl
              << "b = " << std::endl << b << std::endl
              << "x = " << std::endl << x << std::endl;
    return 0;
 }

But there are errors as follow.

main.cpp: In function 'int main()':

main.cpp:23:26: error: invalid initialization of reference of type 'const octave_value_list&' from expression of type 'Matrix'

In file included from main.cpp:3:0:

/usr/local/octave/3.8.0/include/octave-3.8.0/octave/../octave/builtin-defun-decls.h:198:1: error: in passing argument 1 of 'octave_value_list Fmldivide(const octave_value_list&, int)'

Any thoughts?

Answer 1

There are two reasons why this is not working:

1. Fmldivide takes an octave_value_list as input

2. Fmldivide returns an octave_value_list but you are declaring a ColumnVector

You can fix the first by:

1. converting your input into a single octave_value and leaving the conversion to octave_value_list to the compiler:

   Fmldivide (octave_value (b), 1);

2. use do_binary_op with the left division operator:

   do_binary_op (octave_value::op_ldiv, b, 1);

You can fix the second by using the column_vector_value method:

1. if you are using Fmldivide you get a list, so you must index the first element first:

   ColumnVector x = Fmldivide (b, 1)(0).column_vector_value ();

2. if you use do_binary_op you will get an octave_value so indexing is unecessary:

   ColumnVector x = do_binary_op (octave_value::op_ldiv, b, 1).column_vector_value ();