public void GetCalibrationStatus(ref CalibrationOutcome calibOutcome, ref double pricingError)
{
calibOutcome = outcome;
VasicekModel m = new VasicekModel(r0, calibratedParams);
pricingError = CalcMeanSquareErrorBetweenModelAndMarket(m);
}
// Calculate difference between observed and model prices
public double CalcMeanSquareErrorBetweenModelAndMarket(VasicekModel m)
{
double meanSqErr = 0;
foreach (ZCBCallOptionMarketData option in marketOptionsList)
{
double bondMaturity = option.bondMaturity;
double optionExercise = option.optionExercise;
double strike = option.strike;
double modelPrice = m.PriceZCBCall(optionExercise, strike, bondMaturity);
double difference = modelPrice - option.marketMidPrice;
meanSqErr += difference * difference;
}
return(meanSqErr);
}
public void SetGuessParameters(double k, double theta, double sigma)
{
VasicekModel m = new VasicekModel(r0, k, theta, sigma);
calibratedParams = m.ConvertVasicekModeCalibrationParamsToArray();
}
public VasicekModel GetCalibratedModel()
{
VasicekModel m = new VasicekModel(r0, calibratedParams);
return(m);
}
// Used by Alglib minimisation algorithm
public void CalibrationObjectiveFunction(double[] paramsArray, ref double func, object obj)
{
VasicekModel m = new VasicekModel(r0, paramsArray);
func = CalcMeanSquareErrorBetweenModelAndMarket(m);
}
public static double VasicekZCBPrice(double maturity, double r0, double k, double theta, double sigma)
{
VasicekModel m = new VasicekModel(r0, k, theta, sigma);
return(m.PriceZCB(maturity));
}
public static object[,] CalibrateVasicekParameters(
double kGuess,
double thetaGuess,
double sigmaGuess,
double r0,
double[] strikesArray,
double[] bondMaturitiesArray,
double[] optionExerciseTimesArray,
double[] observedPricesArray,
double accuracy,
int maxIterations)
{
if (strikesArray.Length != bondMaturitiesArray.Length ||
bondMaturitiesArray.Length != optionExerciseTimesArray.Length ||
optionExerciseTimesArray.Length != observedPricesArray.Length)
{
throw new ArgumentException("CalibrateVasicekParameters: strikes, optionExerciseTimes, bondMaturities and observedPrices must be of same length.");
}
VasicekCalibrator vasicekCalibrator = new VasicekCalibrator(r0, accuracy, maxIterations);
vasicekCalibrator.SetGuessParameters(kGuess, thetaGuess, sigmaGuess);
int numObservedOptions = strikesArray.Length;
for (int optionIdx = 0; optionIdx < numObservedOptions; ++optionIdx)
{
vasicekCalibrator.AddObservedOption(bondMaturitiesArray[optionIdx],
optionExerciseTimesArray[optionIdx],
strikesArray[optionIdx],
observedPricesArray[optionIdx]);
}
vasicekCalibrator.Calibrate();
CalibrationOutcome outcome = CalibrationOutcome.NotStarted;
double calibrationError = 0;
vasicekCalibrator.GetCalibrationStatus(ref outcome, ref calibrationError);
VasicekModel calibratedModel = vasicekCalibrator.GetCalibratedModel();
// for output
const int numCols = 2;
const int numRows = 5;
object[,] output = new object[numRows, numCols];
output[0, 0] = "k"; output[0, 1] = calibratedModel.GetK();
output[1, 0] = "theta"; output[1, 1] = calibratedModel.GetTheta();
output[2, 0] = "sigma"; output[2, 1] = calibratedModel.GetSigma();
output[3, 0] = "Minimizer Status";
if (outcome == CalibrationOutcome.FinishedOK)
{
output[3, 1] = "OK";
}
else if (outcome == CalibrationOutcome.FailedMaxItReached)
{
output[3, 1] = "Reached max. num. iterations.";
}
else if (outcome == CalibrationOutcome.FailedOtherReason)
{
output[3, 1] = "Failed.";
}
else
{
output[3, 1] = "Unknown outcome.";
}
{
output[4, 0] = "Pricing error"; output[4, 1] = Math.Sqrt(calibrationError);
}
return(output);
}
public static double VasicekZCBCallPrice(double bondMaturity, double optionExercise, double strike, double r0, double k, double theta, double sigma)
{
VasicekModel m = new VasicekModel(r0, k, theta, sigma);
return(m.PriceZCBCall(optionExercise, strike, bondMaturity));
}
public VasicekModel(VasicekModel otherModel) :
this(otherModel.r0, otherModel.k, otherModel.theta, otherModel.sigma)
{
}
public VasicekModel(VasicekModel otherModel)
{
r0 = otherModel.r0; k = otherModel.k; theta = otherModel.theta; sigma = otherModel.sigma;
}