/// <summary>
/// GetSDTMultiplicityResult() calculates Singles, Doubles, and Triples rates plus related metadata.
///
/// This utility method works for both Slow- and Fast-Background Multiplicity Analyzers.
///
/// Note that the input parameter "accidentalsHistogram" need not be normalized,
/// and that a normalized-accidentals distribution will be included in output and will be valid
/// for both slow and fast accidentals.
///
/// </summary>
/// <param name="realsPlusAccidentalsHistogram"> The histogram of gate populations. </param>
/// <param name="accidentalsHistogram"> The histogram of gate populations - NOT NORMALIZED. </param>
/// <param name="wasFastAccidentals"> true or false. Affects how PTsingles are calculated, etc. </param>
/// <param name="multiplicityGateWidth"> as a UInt64, in 100-nanosecond tics. </param>
/// <param name="multiplicityDeadDelay"> as a UInt64, in 100-nanosecond tics. </param>
/// <param name="accidentalsDelay"> as a UInt64, in 100-nanosecond tics. </param>
/// <param name="deadTimeCoeffTinNanoSecs"> as a double, in nanoseconds. </param>
/// <param name="deadTimeCoeffAinMicroSecs"> as a double, in microseconds. </param>
/// <param name="deadTimeCoeffBinPicoSecs"> as a double, in picoseconds. </param>
/// <param name="deadTimeCoeffCinNanoSecs"> as a double, in nanoseconds. </param>
/// <param name="totalMeasurementTime"> as a double, in seconds. </param>
/// <param name="normedAccidentalsHistogram"> UInt64[]. </param>
/// <returns></returns>
public MultiplicityResult GetSDTMultiplicityResult(UInt64[] realsPlusAccidentalsHistogram,
UInt64[] accidentalsHistogram,
Boolean wasFastAccidentals,
UInt64 multiplicityGateWidth,
UInt64 multiplicityDeadDelay,
UInt64 accidentalsDelay,
double deadTimeCoeffTinNanoSecs,
double deadTimeCoeffAinMicroSecs,
double deadTimeCoeffBinPicoSecs,
double deadTimeCoeffCinNanoSecs,
double totalMeasurementTime,
UInt64[] normedAccidentalsHistogram = null)
{
MultiplicityResult result;
double phi;
double gateInSeconds;
UInt32 biggestKey, biggestAKey;
int arrayLength;
double[] alpha;
double[] beta;
//todo: check this out. Makes it slow and is dumb HN -- yes, I said it
BigFloat[] α = new BigFloat[0], β = new BigFloat[0];
result = new MultiplicityResult();
if (wasFastAccidentals == true)
{
result.isSlowBackground = false;
}
else
{
result.isSlowBackground = true;
}
//store parameters
result.multiplicityGateWidth = multiplicityGateWidth;
result.multiplicityDeadDelay = multiplicityDeadDelay;
result.accidentalsDelay = accidentalsDelay;
result.deadTimeCoefficientTinNanoSecs = deadTimeCoeffTinNanoSecs;
result.deadTimeCoefficientAinMicroSecs = deadTimeCoeffAinMicroSecs;
result.deadTimeCoefficientBinPicoSecs = deadTimeCoeffBinPicoSecs;
result.deadTimeCoefficientCinNanoSecs = deadTimeCoeffCinNanoSecs;
//copy the real-plus-accidental multiplicity histogram
biggestKey = 0;
arrayLength = realsPlusAccidentalsHistogram.Length;
for (int i = 0; i < arrayLength; i++)
{
if (realsPlusAccidentalsHistogram[i] > 0)
{
result.realPlusAccidentalDistribution.Add((UInt64)i, realsPlusAccidentalsHistogram[i]);
biggestKey = (UInt32)i;
}
}
result.maxRABin = biggestKey;
//copy the accidental-only histogram
biggestAKey = 0;
arrayLength = accidentalsHistogram.Length;
for (int i = 0; i < arrayLength; i++)
{
if (accidentalsHistogram[i] > 0)
{
result.accidentalDistribution.Add((UInt32)i, accidentalsHistogram[i]);
biggestAKey = (UInt32)i;
}
}
result.maxABin = biggestAKey;
//************************************************************
//Normalize the AccidentalDistribution,
//scaling the FastBackgroundAnalysis result in proportion
//to the number of Real+Accidental gate
UInt64 numAccidentalGates = 0;
UInt64 numRealPlusAccidentalGates = 0;
foreach (KeyValuePair <UInt64, UInt64> pair in result.realPlusAccidentalDistribution)
{
numRealPlusAccidentalGates += pair.Value;
}
// if normed distro presented, compute the normalization param and recompute the unnormalized array from it (JFL but does not seem to work)
if (normedAccidentalsHistogram != null)
{
for (int i = 0; i < normedAccidentalsHistogram.Length; i++)
{
if (normedAccidentalsHistogram[i] > 0)
{
result.normalizedAccidentalDistribution.Add((UInt32)i, normedAccidentalsHistogram[i]);
}
}
UInt64 numNAccidentalGates = 0;
foreach (ulong no in normedAccidentalsHistogram)
{
numNAccidentalGates += no;
}
double denormalizingRatio;
UInt64 denormalizedRate;
denormalizingRatio = ((double)numNAccidentalGates) / ((double)numRealPlusAccidentalGates);
result.accidentalDistribution.Clear();
foreach (KeyValuePair <UInt64, UInt64> pair in result.normalizedAccidentalDistribution)
{
denormalizedRate = (UInt64)(pair.Value * denormalizingRatio);
result.accidentalDistribution.Add(pair.Key, denormalizedRate);
biggestAKey = (UInt32)pair.Key;
}
result.maxABin = biggestAKey;
foreach (KeyValuePair <UInt64, UInt64> pair in result.accidentalDistribution)
{
numAccidentalGates += pair.Value;
}
}
else
{
foreach (KeyValuePair <UInt64, UInt64> pair in result.accidentalDistribution)
{
numAccidentalGates += pair.Value;
}
double normalizingRatio;
UInt64 normalizedRate;
normalizingRatio = ((double)numRealPlusAccidentalGates) / ((double)numAccidentalGates);
result.normalizedAccidentalDistribution.Clear();
foreach (KeyValuePair <UInt64, UInt64> pair in result.accidentalDistribution)
{
normalizedRate = (UInt64)(pair.Value * normalizingRatio);
result.normalizedAccidentalDistribution.Add(pair.Key, normalizedRate);
}
}
//*** END Normalizing the AccidentalDistribution *************
//store the bigger key...
if (biggestAKey > biggestKey)
{
biggestKey = biggestAKey;
}
if (biggestKey < 2)
{
biggestKey = 2; //...minimum size for data-output arrays...
}
alpha = new double[biggestKey + 1];
beta = new double[biggestKey + 1];
gateInSeconds = ((double)multiplicityGateWidth) * ticSizeInSeconds;
phi = (deadTimeCoeffTinNanoSecs / 1E9) / gateInSeconds;
bool standard = true; // dev note: this toggle signals use of BigNum when FP overflow is detected
int axover = 0, bxover = 0;
//calculate the alphas
alpha[0] = 0.0;
alpha[1] = 1.0;
//todo: do math simplification here as done for SR. HN
for (int n = 2; n <= biggestKey; n++)
{
if (phi > 1e-20)
{
alpha[n] = 1.0;
if (standard)
{
for (int k = 0; k <= (n - 2); k++)
{
double alphaCoeff;
alphaCoeff = binomialCoefficient((n - 1), (k + 1))
* Math.Pow((double)(k + 1), (double)k)
* Math.Pow(phi, (double)k)
/ (Math.Pow((1.0 - ((k + 1) * phi)), (double)(k + 2)));
alpha[n] += alphaCoeff;
if (Double.IsInfinity(alpha[n]) || Double.IsNaN(alpha[n]))
{
result.warnings.Add("Overflow alpha at n = " + n + ", k = " + k);
alpha[n] = 0;
axover = n;
standard = false; k = n; n = n - 1; // redo the loop
α = new BigFloat[biggestKey + 1];
Log.TraceEvent(LogLevels.Warning, 13, result.warnings[result.warnings.Count - 1]);
}
}
}
else
{
// INCCCycleConditioning.calc_alpha_beta
PrecisionSpec ps128 = new PrecisionSpec(128, PrecisionSpec.BaseType.BIN);
BigFloat one = new BigFloat(1, ps128);
BigFloat combination;
BigFloat sum;
double raise1, power1, power2;
double log1, log2, log3;
BigFloat exp1, exp2, exp3;
/* calculate alpha array */
sum = new BigFloat(0, ps128);
for (int k = 0; k <= (n - 2); k++)
{
combination = new BigFloat(binomialCoefficient((n - 1), (k + 1)), ps128);
raise1 = (double)(k + 1);
power1 = (double)k;
power2 = (double)(k + 2);
log1 = Math.Log(raise1);
log2 = Math.Log(phi);
log3 = Math.Log(1.0 - raise1 * phi);
exp1 = BigFloat.Exp(new BigFloat(log1 * power1, ps128));
exp2 = BigFloat.Exp(new BigFloat(log2 * power1, ps128));
exp3 = BigFloat.Exp(new BigFloat(log3 * power2, ps128));
sum += combination * exp1 * exp2 / exp3;
}
α[n] = new BigFloat(one + sum, ps128);
}
}
else
{
alpha[n] = n; // Per Daniela 2.22.2018
}
}
//calculate the betas
standard = true;
beta[0] = 0.0;
beta[1] = 0.0;
beta[2] = alpha[2] - 1.0;
for (int n = 3; n <= biggestKey; n++)
{
if (phi > 1e-20)
{
beta[n] = alpha[n] - 1.0;
if (standard)
{
for (int k = 0; k <= (n - 3); k++)
{
double betaCoeff;
betaCoeff = binomialCoefficient((n - 1), (k + 2))
* (k + 1)
* Math.Pow((double)(k + 2), (double)k)
* Math.Pow(phi, (double)k)
/ (Math.Pow((1.0 - ((k + 2) * phi)), (double)(k + 3)));
beta[n] += betaCoeff;
if (Double.IsInfinity(beta[n]) || Double.IsNaN(beta[n]))
{
result.warnings.Add("Overflow beta at n = " + n + ", k = " + k);
beta[n] = 0;
bxover = n;
standard = false; k = n; n = n - 1; // redo the loop
β = new BigFloat[biggestKey + 1];
Log.TraceEvent(LogLevels.Warning, 13, result.warnings[result.warnings.Count - 1]);
}
}
}
else
{
PrecisionSpec ps128 = new PrecisionSpec(128, PrecisionSpec.BaseType.BIN);
BigFloat one = new BigFloat(1, ps128);
BigFloat combination;
BigFloat sum;
double raise1, power1, power2;
double log1, log2, log3;
BigFloat exp1, exp2, exp3;
sum = new BigFloat(0, ps128);
for (int k = 0; k <= n - 3; k++)
{
combination = new BigFloat(binomialCoefficient((n - 1), (k + 2)), ps128);
raise1 = (double)(k + 2);
power1 = (double)k;
power2 = (double)(k + 3);
log1 = Math.Log(raise1);
log2 = Math.Log(phi);
log3 = Math.Log(1.0 - raise1 * phi);
exp1 = BigFloat.Exp(new BigFloat(log1 * power1, ps128));
exp2 = BigFloat.Exp(new BigFloat(log2 * power1, ps128));
exp3 = BigFloat.Exp(new BigFloat(log3 * power2, ps128));
sum += combination * (new BigFloat(k + 1, ps128)) * exp1 * exp2 / exp3;
}
β[n] = α[n] - one + sum;
}
}
else
{
beta[n] = (n * (n - 1) / 2);// Per Daniela 2.22.2018
}
}
//store the alpha and beta coefficients
result.alpha = new double[biggestKey + 1];
result.beta = new double[biggestKey + 1];
for (int i = 0; i <= biggestKey; i++)
{
result.alpha[i] = alpha[i];
result.beta[i] = beta[i];
}
double lastGoodD = 0;
for (int i = axover; axover > 0 && i <= biggestKey; i++)
{
double d;
bool good = Double.TryParse(α[i].ToString(), System.Globalization.NumberStyles.Any, System.Globalization.CultureInfo.InvariantCulture.NumberFormat, out d); // what to do when it is really larger than a double?
if (!good)
{
result.alpha[i] = lastGoodD;
result.warnings.Add(String.Format("α[{0}] conversion failed on {1}", i, α[i].ToString()));
Log.TraceEvent(LogLevels.Warning, 13, result.warnings[result.warnings.Count - 1]);
}
else
{
lastGoodD = d;
result.alpha[i] = d;
}
}
for (int i = bxover; bxover > 0 && i <= biggestKey; i++)
{
double d;
bool good = Double.TryParse(β[i].ToString(), System.Globalization.NumberStyles.Any, System.Globalization.CultureInfo.InvariantCulture.NumberFormat, out d);
if (!good)
{
result.beta[i] = lastGoodD;
result.warnings.Add(String.Format("β[{0}] conversion failed on {1}", i, β[i].ToString())); // URGENT: alpha/beta arrays are to be transformed from doubles to BigFloat types when this conversion happens to overflow
Log.TraceEvent(LogLevels.Warning, 13, result.warnings[result.warnings.Count - 1]);
}
else
{
lastGoodD = d;
result.beta[i] = d;
}
}
//NOTE: in the following calculations,
//variables named RAxxx refer to "Reals Plus Accidentals" phenomena, and
//variables named Axxx refer to "Accidentals" phenomena (such as, fast-background counting)
//calculate the factorial moments
double RAfactorialMoment0, RAfactorialMoment1, RAfactorialMoment2, RAfactorialMoment3;
double AfactorialMoment0, AfactorialMoment1, AfactorialMoment2, AfactorialMoment3;
double RAfactorialMomentAlpha1, AfactorialMomentAlpha1;
double RAfactorialMomentBeta2, AfactorialMomentBeta2;
RAfactorialMoment0 = 0.0;
RAfactorialMoment1 = 0.0;
RAfactorialMoment2 = 0.0;
RAfactorialMoment3 = 0.0;
AfactorialMoment0 = 0.0;
AfactorialMoment1 = 0.0;
AfactorialMoment2 = 0.0;
AfactorialMoment3 = 0.0;
RAfactorialMomentAlpha1 = 0.0;
AfactorialMomentAlpha1 = 0.0;
RAfactorialMomentBeta2 = 0.0;
AfactorialMomentBeta2 = 0.0;
//penultimately, use all this to calculate the SDT rates
double PTsingles = 0;
double RTsingles = 0.0;
double normRAfactorialMoment1 = 0;
double normRAfactorialMoment2 = 0.0;
double normRAfactorialMomentAlpha1 = 0.0;
double normRAfactorialMomentBeta2 = 0.0;
double normAfactorialMoment0 = 0;
double normAfactorialMoment1 = 0.0;
double normAfactorialMoment2 = 0.0;
double normAfactorialMoment3 = 0.0;
double normAfactorialMomentAlpha1 = 0.0;
double normAfactorialMomentBeta2 = 0.0;
for (int i = 0; i <= biggestKey; i++)
{
UInt64 gateCount;
int j = i + 1;
int k = i + 2;
int L = i + 3;
if (result.realPlusAccidentalDistribution.TryGetValue((UInt64)i, out gateCount))
{
RAfactorialMoment0 += (double)gateCount;
}
if (result.accidentalDistribution.TryGetValue((UInt64)i, out gateCount))
{
AfactorialMoment0 += gateCount;
}
if (j <= biggestKey)
{
if (result.realPlusAccidentalDistribution.TryGetValue((UInt64)j, out gateCount))
{
RAfactorialMoment1 += (double)(((UInt64)j) * gateCount);
RAfactorialMomentAlpha1 += alpha[j] * ((double)gateCount);
}
if (result.accidentalDistribution.TryGetValue((UInt64)j, out gateCount))
{
AfactorialMoment1 += (double)(((UInt64)j) * gateCount);
AfactorialMomentAlpha1 += alpha[j] * ((double)gateCount);
}
}
if (k <= biggestKey)
{
if (result.realPlusAccidentalDistribution.TryGetValue((UInt32)k, out gateCount))
{
RAfactorialMoment2 += (double)(((UInt64)(k * (k - 1) / 2)) * gateCount);
RAfactorialMomentBeta2 += beta[k] * ((double)gateCount);
}
if (result.accidentalDistribution.TryGetValue((UInt32)k, out gateCount))
{
AfactorialMoment2 += (double)(((UInt64)(k * (k - 1) / 2)) * gateCount);
AfactorialMomentBeta2 += beta[k] * ((double)gateCount);
}
}
if (L <= biggestKey)
{
if (result.realPlusAccidentalDistribution.TryGetValue((UInt32)L, out gateCount))
{
RAfactorialMoment3 += ((double)(((UInt64)(L * (L - 1) * (L - 2))) * gateCount)) / 6.0;
}
if (result.accidentalDistribution.TryGetValue((UInt32)L, out gateCount))
{
AfactorialMoment3 += ((double)(((UInt64)(L * (L - 1) * (L - 2))) * gateCount)) / 6.0;
}
}
}
//NOTE: in the following calculations,
//variables named PTxxx refer to "Pulse Triggered" phenomena, and
//variables named RTxxx refer to "Regularly Triggered" phenomena (such as, fast-background counting)
if (wasFastAccidentals)
{
PTsingles = RAfactorialMoment0;
//double gateFactor = numAccidentalGates / Math.Floor(totalMeasurementTime / (multiplicityGateWidth * ticSizeInSeconds));
RTsingles = AfactorialMoment1 / (AfactorialMoment0 * TimeSpan.FromTicks((long)multiplicityGateWidth).TotalSeconds);
normRAfactorialMoment1 = RAfactorialMoment1 / PTsingles;
normRAfactorialMoment2 = RAfactorialMoment2 / PTsingles;
//NOT USED: double normRAfactorialMoment3 = RAfactorialMoment3 / PTsingles;
normRAfactorialMomentAlpha1 = RAfactorialMomentAlpha1 / PTsingles;
normRAfactorialMomentBeta2 = RAfactorialMomentBeta2 / PTsingles;
normAfactorialMoment0 = AfactorialMoment0 / AfactorialMoment0;
normAfactorialMoment1 = AfactorialMoment1 / AfactorialMoment0;
normAfactorialMoment2 = AfactorialMoment2 / AfactorialMoment0;
normAfactorialMoment3 = AfactorialMoment3 / AfactorialMoment0;
normAfactorialMomentAlpha1 = AfactorialMomentAlpha1 / AfactorialMoment0;
normAfactorialMomentBeta2 = AfactorialMomentBeta2 / AfactorialMoment0;
}
else
{
PTsingles = RAfactorialMoment0; //XXX SHOULDN'T THIS BE RAfactorialMoment0 not AfactorialMoment0???, answer, no, the two values should be the same, RA and A of 0 are identical for "slow"
RTsingles = AfactorialMoment0;
normRAfactorialMoment1 = RAfactorialMoment1 / PTsingles;
normRAfactorialMoment2 = RAfactorialMoment2 / PTsingles;
//NOT USED: double normRAfactorialMoment3 = RAfactorialMoment3 / PTsingles;
normRAfactorialMomentAlpha1 = RAfactorialMomentAlpha1 / PTsingles;
normRAfactorialMomentBeta2 = RAfactorialMomentBeta2 / PTsingles;
normAfactorialMoment0 = AfactorialMoment0 / RTsingles;
normAfactorialMoment1 = AfactorialMoment1 / RTsingles;
normAfactorialMoment2 = AfactorialMoment2 / RTsingles;
normAfactorialMoment3 = AfactorialMoment3 / RTsingles;
normAfactorialMomentAlpha1 = AfactorialMomentAlpha1 / RTsingles;
normAfactorialMomentBeta2 = AfactorialMomentBeta2 / RTsingles;
}
double RTdoubles = (0.5 / (multiplicityGateWidth * ticSizeInSeconds)) * ((2.0 * normAfactorialMoment2) - Math.Pow(normAfactorialMoment1, 2.0));
double RTtriples = (0.16667 / (multiplicityGateWidth * ticSizeInSeconds))
* ((6.0 * normAfactorialMoment3) - (6.0 * normAfactorialMoment1 * normAfactorialMoment2) + (2.0 * Math.Pow(normAfactorialMoment1, 3)));
double PTdoubles = PTsingles * (normRAfactorialMoment1 - normAfactorialMoment1);
double PTtriples;
double PTtriplesDTCorrected;
//store the factorial moments
result.RAfactorialMoment0 = RAfactorialMoment0;
result.RAfactorialMoment1 = normRAfactorialMoment1;
result.RAfactorialMoment2 = normRAfactorialMoment2;
result.RAfactorialMoment3 = RAfactorialMoment3;
result.AfactorialMoment0 = normAfactorialMoment0;
result.AfactorialMoment1 = normAfactorialMoment1;
result.AfactorialMoment2 = normAfactorialMoment2;
result.AfactorialMoment3 = normAfactorialMoment3;
//Store alphabeta intermediates
result.AfactorialAlphaMoment1 = normAfactorialMomentAlpha1;
result.RAfactorialAlphaMoment1 = normRAfactorialMomentAlpha1;
result.AfactorialBetaMoment2 = normAfactorialMomentBeta2;
result.RAfactorialBetaMoment2 = normRAfactorialMomentBeta2;
if (AfactorialMoment0 != 0.0)
{
PTtriples = PTsingles * ((normRAfactorialMoment2 - normAfactorialMoment2)
- ((normAfactorialMoment1 / normAfactorialMoment0)
* (normRAfactorialMoment1 - normAfactorialMoment1)));
PTtriplesDTCorrected = PTsingles * ((normRAfactorialMomentBeta2 - normAfactorialMomentBeta2)
- ((normAfactorialMomentAlpha1 / normAfactorialMoment0)
* (normRAfactorialMomentAlpha1 - normAfactorialMomentAlpha1)));
}
else
{
PTtriples = 0.0;
PTtriplesDTCorrected = 0.0;
}
if (totalMeasurementTime > 1E-12)
{
PTsingles /= totalMeasurementTime;
PTdoubles /= totalMeasurementTime;
PTtriples /= totalMeasurementTime;
PTtriplesDTCorrected /= totalMeasurementTime;
}
else
{
PTsingles = 0.0;
PTdoubles = 0.0;
PTtriples = 0.0;
PTtriplesDTCorrected = 0.0;
}
result.TotalMeasurementTime = totalMeasurementTime;
result.TotalMeasurementTics = totalMeasurementTime * 10000000;
//store the SDT rates
result.singlesRatePerSecond = PTsingles;
result.doublesRatePerSecond = PTdoubles;
result.triplesRatePerSecond = PTtriples;
//now that rates are calculated, calculate the dead-time corrections
// dead time correction coefficients for RT rates (INCC as well as H&C)
// determined experimentally using D/T ratio - see analysisComparison/triplesDeadTime.xls
// the best fit (poly3) used to reproduce the trend
// note: valid only for sources in the range of ~100n/s - ~500,000 n/s
/** NOT USED ***
* double DTcoeffT0_RT = 3.42854465;
* double DTcoeffT1_RT = 3.35351651E-6;
* double DTcoeffT2_RT = -5.83706327E-12;
* double DTcoeffT3_RT = 2.03604973E-17;
***************/
// dead time correction coefficients for PT rates with background calculated using H&C consecutive gates
// determined experimentally using D/T ratio - see analysisComparison/triplesDeadTime.xls
// the best fit (poly3) used to reproduce the trend
// note: valid only for sources in the range of ~100n/s - ~500,000 n/s
/** NOT USED ***
* double DTcoeffT0_PT = 2.78760077;
* double DTcoeffT1_PT = 2.86078894E-6;
* double DTcoeffT2_PT = -8.21994836E-12;
* double DTcoeffT3_PT = 9.45195862E-17;
***************/
/** NOT USED ***
* double DTcoeffA_RT = 0.2063; // these values were determined using two source method
* double DTcoeffB_RT = 0.04256;
***************/
double DTcoeffA = deadTimeCoeffAinMicroSecs;
double DTcoeffB = deadTimeCoeffBinPicoSecs;
double DTcoeffC = deadTimeCoeffCinNanoSecs;
double DTcoeffT = deadTimeCoeffTinNanoSecs;
double exponent = ((DTcoeffA / 1E6) + ((DTcoeffB / 1E12) * PTsingles)) * PTsingles;
double PTsinglesDTcorr = PTsingles * Math.Exp(exponent / 4.0);
double PTdoublesDTcorr = PTdoubles * Math.Exp(exponent);
double PTtriplesDTcorr = PTtriplesDTCorrected * Math.Exp((DTcoeffT / 1E9) * PTsingles);
/** NOT USED ***
* double RTsinglesDTcorr = RTsingles * Math.Exp(( ((DTcoeffA/1E6) + ((DTcoeffB/1E12)*RTsingles)) *RTsingles)/4.0);
* double RTdoublesDTcorr = RTdoubles * Math.Exp( ((DTcoeffA_RT/1E6) + ((DTcoeffB_RT/1E12)*RTsingles)) *RTsingles);
* double RTtriplesDTcorr = RTtriples* (( (DTcoeffT3_RT*Math.Pow(RTsingles,3))
+ (DTcoeffT2_RT*Math.Pow(RTsingles,2))
+ (DTcoeffT1_RT*RTsingles)
+ DTcoeffT0_RT)
+ /DTcoeffT0_RT);
***************/
//store the dead-time corrected values
result.deadTimeCorrectedSinglesRate = PTsinglesDTcorr;
result.deadTimeCorrectedDoublesRate = PTdoublesDTcorr;
result.deadTimeCorrectedTriplesRate = PTtriplesDTcorr;
//Calculate the Dytlewski Dead-Time-Corrected Rates, based upon
//N. Dytlewski, Dead-time corrections for multiplicity counters,
//Nuclear Instruments and Methods in Physics Research A305(1991)492-494
double P03, P02, P13, P12, P23;
P03 = Math.Pow((1.0 - (2.0 * phi)), 3.0);
P02 = Math.Pow((1.0 - phi), 2.0);
P13 = (2.0 * Math.Pow((1.0 - phi), 3.0)) - (2.0 * P03);
P12 = 1.0 - P02;
P23 = 1.0 + P03 - (2.0 * Math.Pow((1.0 - phi), 3.0));
result.dytlewskiDeadTimeCorrectedTriplesRate = PTtriples / P03;
//Martyn made me do it. hn 2.6.2015
//result.deadTimeCorrectedTriplesRate = result.dytlewskiDeadTimeCorrectedTriplesRate;
result.dytlewskiDeadTimeCorrectedDoublesRate = (PTdoubles / P02) + ((PTtriples * P13) / (P02 * P03));
result.dytlewskiDeadTimeCorrectedSinglesRate = PTsingles + ((P12 * PTdoubles) / P02) + (PTtriples * (((P12 * P13) / (P02 * P03)) - (P23 / P03)));
return(result);
}
/// <summary>
/// The calculate.
/// </summary>
/// <param name="relationshipType">
/// The relationship type.
/// </param>
/// <param name="columnIsPrimaryKey">
/// The column is primary key.
/// </param>
/// <param name="relatedColumnIsPrimaryKey">
/// The related column is primary key.
/// </param>
/// <param name="columnIsForeignKey">
/// The column is foreign key.
/// </param>
/// <param name="columnIsNullable">
/// The column is nullable.
/// </param>
/// <param name="foreignColumnIsNullable">
/// The foreign column is nullable.
/// </param>
/// <param name="relatedColumnIsForeignKey">
/// The related Column Is Foreign Key.
/// </param>
/// <returns>
/// The <see cref="MultiplicityResult"/>.
/// </returns>
/// <exception cref="Exception">
/// </exception>
public static MultiplicityResult Calculate(
RelationshipType relationshipType,
bool columnIsPrimaryKey,
bool relatedColumnIsPrimaryKey,
bool columnIsForeignKey,
bool columnIsNullable,
bool foreignColumnIsNullable,
bool relatedColumnIsForeignKey)
{
MultiplicityResult result = new MultiplicityResult();
if (relationshipType == RelationshipType.ForeignKey)
{
if (columnIsPrimaryKey && relatedColumnIsPrimaryKey)
{
// Extending Tables. i.e) Book --> Product
if (columnIsPrimaryKey && columnIsForeignKey)
{
result.Multiplicity = RelationshipMultiplicity.ZeroToOne;
result.ReferencedMultiplicity = RelationshipMultiplicity.One;
}
else
{
throw new Exception();
}
}
else if (!columnIsPrimaryKey && columnIsForeignKey)
{
result.Multiplicity = RelationshipMultiplicity.Many;
if (columnIsNullable)
{
// Zero to Many:
result.ReferencedMultiplicity = RelationshipMultiplicity.ZeroToOne;
}
else
{
// 1 to Many
result.ReferencedMultiplicity = RelationshipMultiplicity.One;
}
}
else
{
throw new Exception();
}
}
else if (relationshipType == RelationshipType.ForeignKeyChild)
{
if (columnIsPrimaryKey && relatedColumnIsPrimaryKey)
{
// Todo: Extending Tables. i.e) Product --> Book
if (columnIsPrimaryKey && !columnIsForeignKey)
{
result.Multiplicity = RelationshipMultiplicity.One;
result.ReferencedMultiplicity = RelationshipMultiplicity.ZeroToOne;
}
else
{
result.Multiplicity = RelationshipMultiplicity.One;
result.ReferencedMultiplicity = RelationshipMultiplicity.One;
}
}
else if (columnIsPrimaryKey && relatedColumnIsForeignKey)
{
result.ReferencedMultiplicity = RelationshipMultiplicity.Many;
if (foreignColumnIsNullable)
{
// Zero to Many:
result.Multiplicity = RelationshipMultiplicity.ZeroToOne;
}
else
{
// 1 to Many
result.Multiplicity = RelationshipMultiplicity.One;
}
}
else
{
throw new Exception();
}
}
return(result);
}
