int main( ){
USING_NAMESPACE_ACADO
// INTRODUCE THE VARIABLES:
// -------------------------
Parameter V ;
Parameter km;
// READ THE MEASUREMENT FROM A DATA FILE:
// --------------------------------------
Matrix m = readFromFile( "michaelis_menten_data.txt" );
// DEFINE A MEASUREMENT FUNCTION:
// ------------------------------
Function h; // the measurement function
int i;
for( i = 0; i < (int) m.getNumRows(); i++ )
h << V*m(i,0)/(km + m(i,0)) - m(i,1);
// DEFINE A PARAMETER ESTIMATION PROBLEM:
// --------------------------------------
NLP nlp;
nlp.minimizeLSQ( h );
nlp.subjectTo( 0.0 <= V <= 2.0 );
nlp.subjectTo( 0.0 <= km <= 2.0 );
// DEFINE AN OPTIMIZATION ALGORITHM AND SOLVE THE ESTIMATION PROBLEM:
// ------------------------------------------------------------------
ParameterEstimationAlgorithm algorithm(nlp);
algorithm.solve();
VariablesGrid parameters;
algorithm.getParameters( parameters );
return 0;
// GET THE VARIANCE COVARIANCE IN THE SOLUTION:
// ---------------------------------------------
Matrix var;
algorithm.getParameterVarianceCovariance( var );
double LSSE = 2.0*algorithm.getObjectiveValue();
double MSE = LSSE/( m.getNumRows() - 2.0 ); // m.getNumRows() == number of measurements
// 2 == number of parameters
var *= MSE; // rescale the variance-covariance with the MSE factor.
// PRINT THE RESULT ON THE TERMINAL:
// -----------------------------------------------------------------------
printf("\n\nResults for the parameters: \n");
printf("-----------------------------------------------\n");
printf(" V = %.3e +/- %.3e \n", parameters(0,0), sqrt( var(0,0) ) );
printf(" km = %.3e +/- %.3e \n", parameters(0,1), sqrt( var(1,1) ) );
printf("-----------------------------------------------\n\n\n");
return 0;
}
