02_Use_ModelEvaluator.cpp

Go to the documentation of this file.

//@HEADER
// *****************************************************************************
//          Tempus: Time Integration and Sensitivity Analysis Package
//
// Copyright 2017 NTESS and the Tempus contributors.
// SPDX-License-Identifier: BSD-3-Clause
// *****************************************************************************
//@HEADER
 
#include <iomanip>
#include <iostream>
#include <stdlib.h>
#include <math.h>
#include "Teuchos_StandardCatchMacros.hpp"
 
#include "Thyra_VectorStdOps.hpp"
#include "Thyra_DefaultSpmdVectorSpace.hpp"
#include "Thyra_DetachedVectorView.hpp"
 
#include "VanDerPol_ModelEvaluator_02.hpp"
 
 
using namespace std;
using Teuchos::RCP;
 
int main(int argc, char *argv[])
{
  bool verbose = true;
  bool success = false;
  try {
    // Construct ModelEvaluator
    Teuchos::RCP<const Thyra::ModelEvaluator<double> >
      model = Teuchos::rcp(new VanDerPol_ModelEvaluator_02<double>());
 
    // Get initial conditions from ModelEvaluator::getNominalValues.
    int n = 0;
    double time = 0.0;
    bool passed = true;   // ICs are considered passed.
    RCP<Thyra::VectorBase<double> > x_n    =
      model->getNominalValues().get_x()->clone_v();
    RCP<Thyra::VectorBase<double> > xDot_n =
      model->getNominalValues().get_x_dot()->clone_v();
 
    // Timestep size
    double finalTime = 2.0;
    int nTimeSteps = 2000;
    const double constDT = finalTime/nTimeSteps;
 
    // Advance the solution to the next timestep.
    cout << n << "  " << time << "  " << get_ele(*(x_n), 0)
                              << "  " << get_ele(*(x_n), 1) << endl;
    while (passed && time < finalTime && n < nTimeSteps) {
 
      // Initialize next time step
      RCP<Thyra::VectorBase<double> > x_np1 = x_n->clone_v(); // at time index n+1
 
      // Set the timestep and time.
      double dt = constDT;
      time = (n+1)*dt;
 
      // For explicit ODE formulation, xDot = f(x, t),
      // xDot is part of the outArgs.
      auto inArgs  = model->createInArgs();
      auto outArgs = model->createOutArgs();
      inArgs.set_t(time);
      inArgs.set_x(x_n);
      inArgs.set_x_dot(Teuchos::null);
      outArgs.set_f(xDot_n);
 
      // Righthand side evaluation and time-derivative at n.
      model->evalModel(inArgs, outArgs);
 
      // Take the timestep - Forward Euler
      Thyra::V_VpStV(x_np1.ptr(), *x_n, dt, *xDot_n);
 
      // Test if solution has passed.
      if ( std::isnan(Thyra::norm(*x_np1)) ) {
        passed = false;
      } else {
        // Promote to next step (n <- n+1).
        Thyra::V_V(x_n.ptr(), *x_np1);
        n++;
      }
 
      // Output
      if ( n%100 == 0 )
        cout << n << "  " << time << "  " << get_ele(*(x_n), 0)
                                  << "  " << get_ele(*(x_n), 1) << endl;
    }
 
    // Test for regression.
    RCP<Thyra::VectorBase<double> > x_regress = x_n->clone_v();
    {
      Thyra::DetachedVectorView<double> x_regress_view(*x_regress);
      x_regress_view[0] = -1.59496108218721311;
      x_regress_view[1] =  0.96359412806611255;
    }
 
    RCP<Thyra::VectorBase<double> > x_error = x_n->clone_v();
    Thyra::V_VmV(x_error.ptr(), *x_n, *x_regress);
    double x_L2norm_error   = Thyra::norm_2(*x_error  );
    double x_L2norm_regress = Thyra::norm_2(*x_regress);
 
    cout << "Relative L2 Norm of the error (regression) = "
         << x_L2norm_error/x_L2norm_regress << endl;
    if ( x_L2norm_error > 1.0e-08*x_L2norm_regress) {
      passed = false;
      cout << "FAILED regression constraint!" << endl;
    }
    if (passed) success = true;
  }
  TEUCHOS_STANDARD_CATCH_STATEMENTS(verbose, std::cerr, success);
 
  if(success)
    cout << "\nEnd Result: Test Passed!" << std::endl;
 
  return ( success ? EXIT_SUCCESS : EXIT_FAILURE );
}