/* (non-Javadoc)
* @see net.finmath.stochastic.RandomVariableInterface#cap(net.finmath.stochastic.RandomVariableInterface)
*/
public RandomVariableInterface cap(RandomVariableInterface randomVariable) {
// Set time of this random variable to maximum of time with respect to which measurability is
// known.
double newTime = Math.max(time, randomVariable.getFiltrationTime());
if (isDeterministic() && randomVariable.isDeterministic()) {
double newValueIfNonStochastic = FastMath.min(valueIfNonStochastic, randomVariable.get(0));
return new RandomVariable(newTime, newValueIfNonStochastic);
} else if (isDeterministic()) return randomVariable.cap(this);
else {
double[] newRealizations = new double[Math.max(size(), randomVariable.size())];
for (int i = 0; i < newRealizations.length; i++)
newRealizations[i] = FastMath.min(realizations[i], randomVariable.get(i));
return new RandomVariable(newTime, newRealizations);
}
}
/**
* Get the lightning ratio ([0-1]).
*
* @param position the satellite's position in the selected frame.
* @param frame in which is defined the position
* @param date the date
* @return lightning ratio
* @exception OrekitException if the trajectory is inside the Earth
*/
public double getLightningRatio(
final Vector3D position, final Frame frame, final AbsoluteDate date) throws OrekitException {
// Compute useful angles
final double[] angle = getEclipseAngles(position, frame, date);
// Sat-Sun / Sat-CentralBody angle
final double sunEarthAngle = angle[0];
// Central Body apparent radius
final double alphaCentral = angle[1];
// Sun apparent radius
final double alphaSun = angle[2];
double result = 1.0;
// Is the satellite in complete umbra ?
if (sunEarthAngle - alphaCentral + alphaSun <= 0.0) {
result = 0.0;
} else if (sunEarthAngle - alphaCentral - alphaSun < 0.0) {
// Compute a lightning ratio in penumbra
final double sEA2 = sunEarthAngle * sunEarthAngle;
final double oo2sEA = 1.0 / (2. * sunEarthAngle);
final double aS2 = alphaSun * alphaSun;
final double aE2 = alphaCentral * alphaCentral;
final double aE2maS2 = aE2 - aS2;
final double alpha1 = (sEA2 - aE2maS2) * oo2sEA;
final double alpha2 = (sEA2 + aE2maS2) * oo2sEA;
// Protection against numerical inaccuracy at boundaries
final double a1oaS = FastMath.min(1.0, FastMath.max(-1.0, alpha1 / alphaSun));
final double aS2ma12 = FastMath.max(0.0, aS2 - alpha1 * alpha1);
final double a2oaE = FastMath.min(1.0, FastMath.max(-1.0, alpha2 / alphaCentral));
final double aE2ma22 = FastMath.max(0.0, aE2 - alpha2 * alpha2);
final double P1 = aS2 * FastMath.acos(a1oaS) - alpha1 * FastMath.sqrt(aS2ma12);
final double P2 = aE2 * FastMath.acos(a2oaE) - alpha2 * FastMath.sqrt(aE2ma22);
result = 1. - (P1 + P2) / (FastMath.PI * aS2);
}
return result;
}
/**
* Create the objects needed for linear transformation.
*
* <p>Each {@link org.orekit.propagation.semianalytical.dsst.utilities.hansenHansenTesseralLinear
* HansenTesseralLinear} uses a fixed value for s and j. Since j varies from -maxJ to +maxJ and s
* varies from -maxDegree to +maxDegree, a 2 * maxDegree + 1 x 2 * maxJ + 1 matrix of objects
* should be created. The size of this matrix can be reduced by taking into account the expression
* (2.7.3-2). This means that is enough to create the objects for positive values of j and all
* values of s.
*
* @param meanOnly create only the objects required for the mean contribution
*/
private void createHansenObjects(final boolean meanOnly) {
// Allocate the two dimensional array
this.hansenObjects = new HansenTesseralLinear[2 * maxDegree + 1][jMax + 1];
if (meanOnly) {
// loop through the resonant orders
for (int m : resOrders) {
// Compute the corresponding j term
final int j = FastMath.max(1, (int) FastMath.round(ratio * m));
// Compute the sMin and sMax values
final int sMin = FastMath.min(maxEccPow - j, maxDegree);
final int sMax = FastMath.min(maxEccPow + j, maxDegree);
// loop through the s values
for (int s = 0; s <= sMax; s++) {
// Compute the n0 value
final int n0 = FastMath.max(FastMath.max(2, m), s);
// Create the object for the pair j,s
this.hansenObjects[s + maxDegree][j] =
new HansenTesseralLinear(maxDegree, s, j, n0, maxHansen);
if (s > 0 && s <= sMin) {
// Also create the object for the pair j, -s
this.hansenObjects[maxDegree - s][j] =
new HansenTesseralLinear(maxDegree, -s, j, n0, maxHansen);
}
}
}
} else {
// create all objects
for (int j = 0; j <= jMax; j++) {
for (int s = -maxDegree; s <= maxDegree; s++) {
// Compute the n0 value
final int n0 = FastMath.max(2, FastMath.abs(s));
this.hansenObjects[s + maxDegree][j] =
new HansenTesseralLinear(maxDegree, s, j, n0, maxHansen);
}
}
}
}
/** {@inheritDoc} */
@Override
public void initialize(final AuxiliaryElements aux, final boolean meanOnly)
throws OrekitException {
// Keplerian period
orbitPeriod = aux.getKeplerianPeriod();
// orbit frame
frame = aux.getFrame();
// Set the highest power of the eccentricity in the analytical power
// series expansion for the averaged high order resonant central body
// spherical harmonic perturbation
final double e = aux.getEcc();
if (e <= 0.005) {
maxEccPow = 3;
} else if (e <= 0.02) {
maxEccPow = 4;
} else if (e <= 0.1) {
maxEccPow = 7;
} else if (e <= 0.2) {
maxEccPow = 10;
} else if (e <= 0.3) {
maxEccPow = 12;
} else if (e <= 0.4) {
maxEccPow = 15;
} else {
maxEccPow = 20;
}
// Set the maximum power of the eccentricity to use in Hansen coefficient Kernel expansion.
maxHansen = maxEccPow / 2;
jMax = FastMath.min(MAXJ, maxDegree + maxEccPow);
// Ratio of satellite to central body periods to define resonant terms
ratio = orbitPeriod / bodyPeriod;
// Compute the resonant tesseral harmonic terms if not set by the user
getResonantAndNonResonantTerms(meanOnly);
// initialize the HansenTesseralLinear objects needed
createHansenObjects(meanOnly);
if (!meanOnly) {
// Initialize the Tesseral Short Periodics coefficient class
tesseralSPCoefs =
new TesseralShortPeriodicCoefficients(
jMax,
FastMath.max(maxOrderTesseralSP, maxOrderMdailyTesseralSP),
INTERPOLATION_POINTS);
}
}
@Test
public void testNextLongWideRange() {
long lower = -0x6543210FEDCBA987L;
long upper = 0x456789ABCDEF0123L;
long max = Long.MIN_VALUE;
long min = Long.MAX_VALUE;
for (int i = 0; i < 10000000; ++i) {
long r = randomData.nextLong(lower, upper);
max = FastMath.max(max, r);
min = FastMath.min(min, r);
Assert.assertTrue(r >= lower);
Assert.assertTrue(r <= upper);
}
double ratio = (((double) max) - ((double) min)) / (((double) upper) - ((double) lower));
Assert.assertTrue(ratio > 0.99999);
}
public static int calculateCityLoyalty(int tax_rate, EfsIni efs_ini, Game game) {
int excom_penalty = 0;
// System.out.println(" state = " + game.getDiplomacy().getDiplomaticState(game.getTurn(),
// C.THE_CHURCH));
if (game.getDiplomacy().getDiplomaticState(game.getTurn(), C.THE_CHURCH) == C.DS_WAR) {
excom_penalty = game.getEfs_ini().excom_peasant_loyalty_hit;
// System.out.println("WTF");
}
// System.out.println("excom_peasant_loyalty_hit = " +
// game.getEfs_ini().excom_peasant_loyalty_hit);
// System.out.println("Excom penalty = " + excom_penalty);
return FastMath.max(
0,
FastMath.min(
100, 100 - (tax_rate - efs_ini.default_tax_rate) * C.TAX_LOYALTY_HIT - excom_penalty));
}
@Test
public void testNextIntWideRange() {
int lower = -0x6543210F;
int upper = 0x456789AB;
int max = Integer.MIN_VALUE;
int min = Integer.MAX_VALUE;
for (int i = 0; i < 1000000; ++i) {
int r = randomData.nextInt(lower, upper);
max = FastMath.max(max, r);
min = FastMath.min(min, r);
Assert.assertTrue(r >= lower);
Assert.assertTrue(r <= upper);
}
double ratio = (((double) max) - ((double) min)) / (((double) upper) - ((double) lower));
Assert.assertTrue(ratio > 0.99999);
}
/**
* Returns the value of Δ(p) + Δ(q) - Δ(p + q), with p, q ≥ 10. Based on the <em>NSWC Library of
* Mathematics Subroutines</em> double precision implementation, {@code DBCORR}. In {@code
* BetaTest.testSumDeltaMinusDeltaSum()}, this private method is accessed through reflection.
*
* @param p First argument.
* @param q Second argument.
* @return the value of {@code Delta(p) + Delta(q) - Delta(p + q)}.
* @throws NumberIsTooSmallException if {@code p < 10.0} or {@code q < 10.0}.
*/
private static double sumDeltaMinusDeltaSum(final double p, final double q) {
if (p < 10.0) {
throw new NumberIsTooSmallException(p, 10.0, true);
}
if (q < 10.0) {
throw new NumberIsTooSmallException(q, 10.0, true);
}
final double a = FastMath.min(p, q);
final double b = FastMath.max(p, q);
final double sqrtT = 10.0 / a;
final double t = sqrtT * sqrtT;
double z = DELTA[DELTA.length - 1];
for (int i = DELTA.length - 2; i >= 0; i--) {
z = t * z + DELTA[i];
}
return z / a + deltaMinusDeltaSum(a, b);
}
/**
* Single constructor.
*
* @param centralBodyFrame rotating body frame
* @param centralBodyRotationRate central body rotation rate (rad/s)
* @param provider provider for spherical harmonics
* @param mDailiesOnly if true only M-dailies tesseral harmonics are taken into account for short
* periodics
*/
TesseralContribution(
final Frame centralBodyFrame,
final double centralBodyRotationRate,
final UnnormalizedSphericalHarmonicsProvider provider,
final boolean mDailiesOnly) {
// Central body rotating frame
this.bodyFrame = centralBodyFrame;
// Save the rotation rate
this.centralBodyRotationRate = centralBodyRotationRate;
// Central body rotation period in seconds
this.bodyPeriod = MathUtils.TWO_PI / centralBodyRotationRate;
// Provider for spherical harmonics
this.provider = provider;
this.maxDegree = provider.getMaxDegree();
this.maxOrder = provider.getMaxOrder();
// set the maximum degree order for short periodics
this.maxDegreeTesseralSP = FastMath.min(maxDegree, MAX_DEGREE_TESSERAL_SP);
this.maxDegreeMdailyTesseralSP = FastMath.min(maxDegree, MAX_DEGREE_MDAILY_TESSERAL_SP);
this.maxOrderTesseralSP = FastMath.min(maxOrder, MAX_ORDER_TESSERAL_SP);
this.maxOrderMdailyTesseralSP = FastMath.min(maxOrder, MAX_ORDER_MDAILY_TESSERAL_SP);
// set the maximum value for eccentricity power
this.maxEccPowTesseralSP = MAX_ECCPOWER_SP;
this.maxEccPowMdailyTesseralSP = FastMath.min(maxDegreeMdailyTesseralSP - 2, MAX_ECCPOWER_SP);
this.jMax = FastMath.min(MAXJ, maxDegreeTesseralSP + maxEccPowTesseralSP);
// m-daylies only
this.mDailiesOnly = mDailiesOnly;
// Initialize default values
this.resOrders = new ArrayList<Integer>();
this.nonResOrders = new TreeMap<Integer, List<Integer>>();
this.maxEccPow = 0;
this.maxHansen = 0;
// Factorials computation
final int maxFact = 2 * maxDegree + 1;
this.fact = new double[maxFact];
fact[0] = 1;
for (int i = 1; i < maxFact; i++) {
fact[i] = i * fact[i - 1];
}
}
/**
* Calculates the compact Singular Value Decomposition of the given matrix.
*
* @param matrix Matrix to decompose.
*/
public SingularValueDecomposition(final RealMatrix matrix) {
final double[][] A;
// "m" is always the largest dimension.
if (matrix.getRowDimension() < matrix.getColumnDimension()) {
transposed = true;
A = matrix.transpose().getData();
m = matrix.getColumnDimension();
n = matrix.getRowDimension();
} else {
transposed = false;
A = matrix.getData();
m = matrix.getRowDimension();
n = matrix.getColumnDimension();
}
singularValues = new double[n];
final double[][] U = new double[m][n];
final double[][] V = new double[n][n];
final double[] e = new double[n];
final double[] work = new double[m];
// Reduce A to bidiagonal form, storing the diagonal elements
// in s and the super-diagonal elements in e.
final int nct = FastMath.min(m - 1, n);
final int nrt = FastMath.max(0, n - 2);
for (int k = 0; k < FastMath.max(nct, nrt); k++) {
if (k < nct) {
// Compute the transformation for the k-th column and
// place the k-th diagonal in s[k].
// Compute 2-norm of k-th column without under/overflow.
singularValues[k] = 0;
for (int i = k; i < m; i++) {
singularValues[k] = FastMath.hypot(singularValues[k], A[i][k]);
}
if (singularValues[k] != 0) {
if (A[k][k] < 0) {
singularValues[k] = -singularValues[k];
}
for (int i = k; i < m; i++) {
A[i][k] /= singularValues[k];
}
A[k][k] += 1;
}
singularValues[k] = -singularValues[k];
}
for (int j = k + 1; j < n; j++) {
if (k < nct && singularValues[k] != 0) {
// Apply the transformation.
double t = 0;
for (int i = k; i < m; i++) {
t += A[i][k] * A[i][j];
}
t = -t / A[k][k];
for (int i = k; i < m; i++) {
A[i][j] += t * A[i][k];
}
}
// Place the k-th row of A into e for the
// subsequent calculation of the row transformation.
e[j] = A[k][j];
}
if (k < nct) {
// Place the transformation in U for subsequent back
// multiplication.
for (int i = k; i < m; i++) {
U[i][k] = A[i][k];
}
}
if (k < nrt) {
// Compute the k-th row transformation and place the
// k-th super-diagonal in e[k].
// Compute 2-norm without under/overflow.
e[k] = 0;
for (int i = k + 1; i < n; i++) {
e[k] = FastMath.hypot(e[k], e[i]);
}
if (e[k] != 0) {
if (e[k + 1] < 0) {
e[k] = -e[k];
}
for (int i = k + 1; i < n; i++) {
e[i] /= e[k];
}
e[k + 1] += 1;
}
e[k] = -e[k];
if (k + 1 < m && e[k] != 0) {
// Apply the transformation.
for (int i = k + 1; i < m; i++) {
work[i] = 0;
}
for (int j = k + 1; j < n; j++) {
for (int i = k + 1; i < m; i++) {
work[i] += e[j] * A[i][j];
}
}
for (int j = k + 1; j < n; j++) {
final double t = -e[j] / e[k + 1];
for (int i = k + 1; i < m; i++) {
A[i][j] += t * work[i];
}
}
}
// Place the transformation in V for subsequent
// back multiplication.
for (int i = k + 1; i < n; i++) {
V[i][k] = e[i];
}
}
}
// Set up the final bidiagonal matrix or order p.
int p = n;
if (nct < n) {
singularValues[nct] = A[nct][nct];
}
if (m < p) {
singularValues[p - 1] = 0;
}
if (nrt + 1 < p) {
e[nrt] = A[nrt][p - 1];
}
e[p - 1] = 0;
// Generate U.
for (int j = nct; j < n; j++) {
for (int i = 0; i < m; i++) {
U[i][j] = 0;
}
U[j][j] = 1;
}
for (int k = nct - 1; k >= 0; k--) {
if (singularValues[k] != 0) {
for (int j = k + 1; j < n; j++) {
double t = 0;
for (int i = k; i < m; i++) {
t += U[i][k] * U[i][j];
}
t = -t / U[k][k];
for (int i = k; i < m; i++) {
U[i][j] += t * U[i][k];
}
}
for (int i = k; i < m; i++) {
U[i][k] = -U[i][k];
}
U[k][k] = 1 + U[k][k];
for (int i = 0; i < k - 1; i++) {
U[i][k] = 0;
}
} else {
for (int i = 0; i < m; i++) {
U[i][k] = 0;
}
U[k][k] = 1;
}
}
// Generate V.
for (int k = n - 1; k >= 0; k--) {
if (k < nrt && e[k] != 0) {
for (int j = k + 1; j < n; j++) {
double t = 0;
for (int i = k + 1; i < n; i++) {
t += V[i][k] * V[i][j];
}
t = -t / V[k + 1][k];
for (int i = k + 1; i < n; i++) {
V[i][j] += t * V[i][k];
}
}
}
for (int i = 0; i < n; i++) {
V[i][k] = 0;
}
V[k][k] = 1;
}
// Main iteration loop for the singular values.
final int pp = p - 1;
int iter = 0;
while (p > 0) {
int k;
int kase;
// Here is where a test for too many iterations would go.
// This section of the program inspects for
// negligible elements in the s and e arrays. On
// completion the variables kase and k are set as follows.
// kase = 1 if s(p) and e[k-1] are negligible and k<p
// kase = 2 if s(k) is negligible and k<p
// kase = 3 if e[k-1] is negligible, k<p, and
// s(k), ..., s(p) are not negligible (qr step).
// kase = 4 if e(p-1) is negligible (convergence).
for (k = p - 2; k >= 0; k--) {
final double threshold =
TINY + EPS * (FastMath.abs(singularValues[k]) + FastMath.abs(singularValues[k + 1]));
// the following condition is written this way in order
// to break out of the loop when NaN occurs, writing it
// as "if (FastMath.abs(e[k]) <= threshold)" would loop
// indefinitely in case of NaNs because comparison on NaNs
// always return false, regardless of what is checked
// see issue MATH-947
if (!(FastMath.abs(e[k]) > threshold)) {
e[k] = 0;
break;
}
}
if (k == p - 2) {
kase = 4;
} else {
int ks;
for (ks = p - 1; ks >= k; ks--) {
if (ks == k) {
break;
}
final double t =
(ks != p ? FastMath.abs(e[ks]) : 0) + (ks != k + 1 ? FastMath.abs(e[ks - 1]) : 0);
if (FastMath.abs(singularValues[ks]) <= TINY + EPS * t) {
singularValues[ks] = 0;
break;
}
}
if (ks == k) {
kase = 3;
} else if (ks == p - 1) {
kase = 1;
} else {
kase = 2;
k = ks;
}
}
k++;
// Perform the task indicated by kase.
switch (kase) {
// Deflate negligible s(p).
case 1:
{
double f = e[p - 2];
e[p - 2] = 0;
for (int j = p - 2; j >= k; j--) {
double t = FastMath.hypot(singularValues[j], f);
final double cs = singularValues[j] / t;
final double sn = f / t;
singularValues[j] = t;
if (j != k) {
f = -sn * e[j - 1];
e[j - 1] = cs * e[j - 1];
}
for (int i = 0; i < n; i++) {
t = cs * V[i][j] + sn * V[i][p - 1];
V[i][p - 1] = -sn * V[i][j] + cs * V[i][p - 1];
V[i][j] = t;
}
}
}
break;
// Split at negligible s(k).
case 2:
{
double f = e[k - 1];
e[k - 1] = 0;
for (int j = k; j < p; j++) {
double t = FastMath.hypot(singularValues[j], f);
final double cs = singularValues[j] / t;
final double sn = f / t;
singularValues[j] = t;
f = -sn * e[j];
e[j] = cs * e[j];
for (int i = 0; i < m; i++) {
t = cs * U[i][j] + sn * U[i][k - 1];
U[i][k - 1] = -sn * U[i][j] + cs * U[i][k - 1];
U[i][j] = t;
}
}
}
break;
// Perform one qr step.
case 3:
{
// Calculate the shift.
final double maxPm1Pm2 =
FastMath.max(
FastMath.abs(singularValues[p - 1]), FastMath.abs(singularValues[p - 2]));
final double scale =
FastMath.max(
FastMath.max(
FastMath.max(maxPm1Pm2, FastMath.abs(e[p - 2])),
FastMath.abs(singularValues[k])),
FastMath.abs(e[k]));
final double sp = singularValues[p - 1] / scale;
final double spm1 = singularValues[p - 2] / scale;
final double epm1 = e[p - 2] / scale;
final double sk = singularValues[k] / scale;
final double ek = e[k] / scale;
final double b = ((spm1 + sp) * (spm1 - sp) + epm1 * epm1) / 2.0;
final double c = (sp * epm1) * (sp * epm1);
double shift = 0;
if (b != 0 || c != 0) {
shift = FastMath.sqrt(b * b + c);
if (b < 0) {
shift = -shift;
}
shift = c / (b + shift);
}
double f = (sk + sp) * (sk - sp) + shift;
double g = sk * ek;
// Chase zeros.
for (int j = k; j < p - 1; j++) {
double t = FastMath.hypot(f, g);
double cs = f / t;
double sn = g / t;
if (j != k) {
e[j - 1] = t;
}
f = cs * singularValues[j] + sn * e[j];
e[j] = cs * e[j] - sn * singularValues[j];
g = sn * singularValues[j + 1];
singularValues[j + 1] = cs * singularValues[j + 1];
for (int i = 0; i < n; i++) {
t = cs * V[i][j] + sn * V[i][j + 1];
V[i][j + 1] = -sn * V[i][j] + cs * V[i][j + 1];
V[i][j] = t;
}
t = FastMath.hypot(f, g);
cs = f / t;
sn = g / t;
singularValues[j] = t;
f = cs * e[j] + sn * singularValues[j + 1];
singularValues[j + 1] = -sn * e[j] + cs * singularValues[j + 1];
g = sn * e[j + 1];
e[j + 1] = cs * e[j + 1];
if (j < m - 1) {
for (int i = 0; i < m; i++) {
t = cs * U[i][j] + sn * U[i][j + 1];
U[i][j + 1] = -sn * U[i][j] + cs * U[i][j + 1];
U[i][j] = t;
}
}
}
e[p - 2] = f;
iter = iter + 1;
}
break;
// Convergence.
default:
{
// Make the singular values positive.
if (singularValues[k] <= 0) {
singularValues[k] = singularValues[k] < 0 ? -singularValues[k] : 0;
for (int i = 0; i <= pp; i++) {
V[i][k] = -V[i][k];
}
}
// Order the singular values.
while (k < pp) {
if (singularValues[k] >= singularValues[k + 1]) {
break;
}
double t = singularValues[k];
singularValues[k] = singularValues[k + 1];
singularValues[k + 1] = t;
if (k < n - 1) {
for (int i = 0; i < n; i++) {
t = V[i][k + 1];
V[i][k + 1] = V[i][k];
V[i][k] = t;
}
}
if (k < m - 1) {
for (int i = 0; i < m; i++) {
t = U[i][k + 1];
U[i][k + 1] = U[i][k];
U[i][k] = t;
}
}
k++;
}
iter = 0;
p--;
}
break;
}
}
// Set the small value tolerance used to calculate rank and pseudo-inverse
tol = FastMath.max(m * singularValues[0] * EPS, FastMath.sqrt(Precision.SAFE_MIN));
if (!transposed) {
cachedU = MatrixUtils.createRealMatrix(U);
cachedV = MatrixUtils.createRealMatrix(V);
} else {
cachedU = MatrixUtils.createRealMatrix(V);
cachedV = MatrixUtils.createRealMatrix(U);
}
}
/** {@inheritDoc} */
@Override
public void integrate(final ExpandableStatefulODE equations, final double t)
throws MathIllegalStateException, MathIllegalArgumentException {
sanityChecks(equations, t);
setEquations(equations);
final boolean forward = t > equations.getTime();
// create some internal working arrays
final double[] y0 = equations.getCompleteState();
final double[] y = y0.clone();
final int stages = c.length + 1;
final double[][] yDotK = new double[stages][y.length];
final double[] yTmp = y0.clone();
final double[] yDotTmp = new double[y.length];
// set up an interpolator sharing the integrator arrays
final RungeKuttaStepInterpolator interpolator = (RungeKuttaStepInterpolator) prototype.copy();
interpolator.reinitialize(
this, yTmp, yDotK, forward, equations.getPrimaryMapper(), equations.getSecondaryMappers());
interpolator.storeTime(equations.getTime());
// set up integration control objects
stepStart = equations.getTime();
double hNew = 0;
boolean firstTime = true;
initIntegration(equations.getTime(), y0, t);
// main integration loop
isLastStep = false;
do {
interpolator.shift();
// iterate over step size, ensuring local normalized error is smaller than 1
double error = 10;
while (error >= 1.0) {
if (firstTime || !fsal) {
// first stage
computeDerivatives(stepStart, y, yDotK[0]);
}
if (firstTime) {
final double[] scale = new double[mainSetDimension];
if (vecAbsoluteTolerance == null) {
for (int i = 0; i < scale.length; ++i) {
scale[i] = scalAbsoluteTolerance + scalRelativeTolerance * FastMath.abs(y[i]);
}
} else {
for (int i = 0; i < scale.length; ++i) {
scale[i] = vecAbsoluteTolerance[i] + vecRelativeTolerance[i] * FastMath.abs(y[i]);
}
}
hNew = initializeStep(forward, getOrder(), scale, stepStart, y, yDotK[0], yTmp, yDotK[1]);
firstTime = false;
}
stepSize = hNew;
if (forward) {
if (stepStart + stepSize >= t) {
stepSize = t - stepStart;
}
} else {
if (stepStart + stepSize <= t) {
stepSize = t - stepStart;
}
}
// next stages
for (int k = 1; k < stages; ++k) {
for (int j = 0; j < y0.length; ++j) {
double sum = a[k - 1][0] * yDotK[0][j];
for (int l = 1; l < k; ++l) {
sum += a[k - 1][l] * yDotK[l][j];
}
yTmp[j] = y[j] + stepSize * sum;
}
computeDerivatives(stepStart + c[k - 1] * stepSize, yTmp, yDotK[k]);
}
// estimate the state at the end of the step
for (int j = 0; j < y0.length; ++j) {
double sum = b[0] * yDotK[0][j];
for (int l = 1; l < stages; ++l) {
sum += b[l] * yDotK[l][j];
}
yTmp[j] = y[j] + stepSize * sum;
}
// estimate the error at the end of the step
error = estimateError(yDotK, y, yTmp, stepSize);
if (error >= 1.0) {
// reject the step and attempt to reduce error by stepsize control
final double factor =
FastMath.min(
maxGrowth, FastMath.max(minReduction, safety * FastMath.pow(error, exp)));
hNew = filterStep(stepSize * factor, forward, false);
}
}
// local error is small enough: accept the step, trigger events and step handlers
interpolator.storeTime(stepStart + stepSize);
System.arraycopy(yTmp, 0, y, 0, y0.length);
System.arraycopy(yDotK[stages - 1], 0, yDotTmp, 0, y0.length);
stepStart = acceptStep(interpolator, y, yDotTmp, t);
System.arraycopy(y, 0, yTmp, 0, y.length);
if (!isLastStep) {
// prepare next step
interpolator.storeTime(stepStart);
if (fsal) {
// save the last evaluation for the next step
System.arraycopy(yDotTmp, 0, yDotK[0], 0, y0.length);
}
// stepsize control for next step
final double factor =
FastMath.min(maxGrowth, FastMath.max(minReduction, safety * FastMath.pow(error, exp)));
final double scaledH = stepSize * factor;
final double nextT = stepStart + scaledH;
final boolean nextIsLast = forward ? (nextT >= t) : (nextT <= t);
hNew = filterStep(scaledH, forward, nextIsLast);
final double filteredNextT = stepStart + hNew;
final boolean filteredNextIsLast = forward ? (filteredNextT >= t) : (filteredNextT <= t);
if (filteredNextIsLast) {
hNew = t - stepStart;
}
}
} while (!isLastStep);
// dispatch results
equations.setTime(stepStart);
equations.setCompleteState(y);
resetInternalState();
}
/**
* Determines the Levenberg-Marquardt parameter.
*
* <p>This implementation is a translation in Java of the MINPACK <a
* href="http://www.netlib.org/minpack/lmpar.f">lmpar</a> routine.
*
* <p>This method sets the lmPar and lmDir attributes.
*
* <p>The authors of the original fortran function are:
*
* <ul>
* <li>Argonne National Laboratory. MINPACK project. March 1980
* <li>Burton S. Garbow
* <li>Kenneth E. Hillstrom
* <li>Jorge J. More
* </ul>
*
* <p>Luc Maisonobe did the Java translation.
*
* @param qy Array containing qTy.
* @param delta Upper bound on the euclidean norm of diagR * lmDir.
* @param diag Diagonal matrix.
* @param internalData Data (modified in-place in this method).
* @param solvedCols Number of solved point.
* @param work1 work array
* @param work2 work array
* @param work3 work array
* @param lmDir the "returned" LM direction will be stored in this array.
* @param lmPar the value of the LM parameter from the previous iteration.
* @return the new LM parameter
*/
private double determineLMParameter(
double[] qy,
double delta,
double[] diag,
InternalData internalData,
int solvedCols,
double[] work1,
double[] work2,
double[] work3,
double[] lmDir,
double lmPar) {
final double[][] weightedJacobian = internalData.weightedJacobian;
final int[] permutation = internalData.permutation;
final int rank = internalData.rank;
final double[] diagR = internalData.diagR;
final int nC = weightedJacobian[0].length;
// compute and store in x the gauss-newton direction, if the
// jacobian is rank-deficient, obtain a least squares solution
for (int j = 0; j < rank; ++j) {
lmDir[permutation[j]] = qy[j];
}
for (int j = rank; j < nC; ++j) {
lmDir[permutation[j]] = 0;
}
for (int k = rank - 1; k >= 0; --k) {
int pk = permutation[k];
double ypk = lmDir[pk] / diagR[pk];
for (int i = 0; i < k; ++i) {
lmDir[permutation[i]] -= ypk * weightedJacobian[i][pk];
}
lmDir[pk] = ypk;
}
// evaluate the function at the origin, and test
// for acceptance of the Gauss-Newton direction
double dxNorm = 0;
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
double s = diag[pj] * lmDir[pj];
work1[pj] = s;
dxNorm += s * s;
}
dxNorm = FastMath.sqrt(dxNorm);
double fp = dxNorm - delta;
if (fp <= 0.1 * delta) {
lmPar = 0;
return lmPar;
}
// if the jacobian is not rank deficient, the Newton step provides
// a lower bound, parl, for the zero of the function,
// otherwise set this bound to zero
double sum2;
double parl = 0;
if (rank == solvedCols) {
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
work1[pj] *= diag[pj] / dxNorm;
}
sum2 = 0;
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
double sum = 0;
for (int i = 0; i < j; ++i) {
sum += weightedJacobian[i][pj] * work1[permutation[i]];
}
double s = (work1[pj] - sum) / diagR[pj];
work1[pj] = s;
sum2 += s * s;
}
parl = fp / (delta * sum2);
}
// calculate an upper bound, paru, for the zero of the function
sum2 = 0;
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
double sum = 0;
for (int i = 0; i <= j; ++i) {
sum += weightedJacobian[i][pj] * qy[i];
}
sum /= diag[pj];
sum2 += sum * sum;
}
double gNorm = FastMath.sqrt(sum2);
double paru = gNorm / delta;
if (paru == 0) {
paru = Precision.SAFE_MIN / FastMath.min(delta, 0.1);
}
// if the input par lies outside of the interval (parl,paru),
// set par to the closer endpoint
lmPar = FastMath.min(paru, FastMath.max(lmPar, parl));
if (lmPar == 0) {
lmPar = gNorm / dxNorm;
}
for (int countdown = 10; countdown >= 0; --countdown) {
// evaluate the function at the current value of lmPar
if (lmPar == 0) {
lmPar = FastMath.max(Precision.SAFE_MIN, 0.001 * paru);
}
double sPar = FastMath.sqrt(lmPar);
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
work1[pj] = sPar * diag[pj];
}
determineLMDirection(qy, work1, work2, internalData, solvedCols, work3, lmDir);
dxNorm = 0;
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
double s = diag[pj] * lmDir[pj];
work3[pj] = s;
dxNorm += s * s;
}
dxNorm = FastMath.sqrt(dxNorm);
double previousFP = fp;
fp = dxNorm - delta;
// if the function is small enough, accept the current value
// of lmPar, also test for the exceptional cases where parl is zero
if (FastMath.abs(fp) <= 0.1 * delta || (parl == 0 && fp <= previousFP && previousFP < 0)) {
return lmPar;
}
// compute the Newton correction
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
work1[pj] = work3[pj] * diag[pj] / dxNorm;
}
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
work1[pj] /= work2[j];
double tmp = work1[pj];
for (int i = j + 1; i < solvedCols; ++i) {
work1[permutation[i]] -= weightedJacobian[i][pj] * tmp;
}
}
sum2 = 0;
for (int j = 0; j < solvedCols; ++j) {
double s = work1[permutation[j]];
sum2 += s * s;
}
double correction = fp / (delta * sum2);
// depending on the sign of the function, update parl or paru.
if (fp > 0) {
parl = FastMath.max(parl, lmPar);
} else if (fp < 0) {
paru = FastMath.min(paru, lmPar);
}
// compute an improved estimate for lmPar
lmPar = FastMath.max(parl, lmPar + correction);
}
return lmPar;
}
/**
* Returns the value of log B(p, q) for 0 ≤ x ≤ 1 and p, q > 0. Based on the <em>NSWC Library of
* Mathematics Subroutines</em> implementation, {@code DBETLN}.
*
* @param p First argument.
* @param q Second argument.
* @return the value of {@code log(Beta(p, q))}, {@code NaN} if {@code p <= 0} or {@code q <= 0}.
*/
public static double logBeta(final double p, final double q) {
if (Double.isNaN(p) || Double.isNaN(q) || (p <= 0.0) || (q <= 0.0)) {
return Double.NaN;
}
final double a = FastMath.min(p, q);
final double b = FastMath.max(p, q);
if (a >= 10.0) {
final double w = sumDeltaMinusDeltaSum(a, b);
final double h = a / b;
final double c = h / (1.0 + h);
final double u = -(a - 0.5) * FastMath.log(c);
final double v = b * FastMath.log1p(h);
if (u <= v) {
return (((-0.5 * FastMath.log(b) + HALF_LOG_TWO_PI) + w) - u) - v;
} else {
return (((-0.5 * FastMath.log(b) + HALF_LOG_TWO_PI) + w) - v) - u;
}
} else if (a > 2.0) {
if (b > 1000.0) {
final int n = (int) FastMath.floor(a - 1.0);
double prod = 1.0;
double ared = a;
for (int i = 0; i < n; i++) {
ared -= 1.0;
prod *= ared / (1.0 + ared / b);
}
return (FastMath.log(prod) - n * FastMath.log(b))
+ (Gamma.logGamma(ared) + logGammaMinusLogGammaSum(ared, b));
} else {
double prod1 = 1.0;
double ared = a;
while (ared > 2.0) {
ared -= 1.0;
final double h = ared / b;
prod1 *= h / (1.0 + h);
}
if (b < 10.0) {
double prod2 = 1.0;
double bred = b;
while (bred > 2.0) {
bred -= 1.0;
prod2 *= bred / (ared + bred);
}
return FastMath.log(prod1)
+ FastMath.log(prod2)
+ (Gamma.logGamma(ared) + (Gamma.logGamma(bred) - logGammaSum(ared, bred)));
} else {
return FastMath.log(prod1) + Gamma.logGamma(ared) + logGammaMinusLogGammaSum(ared, b);
}
}
} else if (a >= 1.0) {
if (b > 2.0) {
if (b < 10.0) {
double prod = 1.0;
double bred = b;
while (bred > 2.0) {
bred -= 1.0;
prod *= bred / (a + bred);
}
return FastMath.log(prod)
+ (Gamma.logGamma(a) + (Gamma.logGamma(bred) - logGammaSum(a, bred)));
} else {
return Gamma.logGamma(a) + logGammaMinusLogGammaSum(a, b);
}
} else {
return Gamma.logGamma(a) + Gamma.logGamma(b) - logGammaSum(a, b);
}
} else {
if (b >= 10.0) {
return Gamma.logGamma(a) + logGammaMinusLogGammaSum(a, b);
} else {
// The following command is the original NSWC implementation.
// return Gamma.logGamma(a) +
// (Gamma.logGamma(b) - Gamma.logGamma(a + b));
// The following command turns out to be more accurate.
return FastMath.log(Gamma.gamma(a) * Gamma.gamma(b) / Gamma.gamma(a + b));
}
}
}
/**
* Compute the n-SUM for potential derivatives components.
*
* @param date current date
* @param j resonant index <i>j</i>
* @param m resonant order <i>m</i>
* @param s d'Alembert characteristic <i>s</i>
* @param maxN maximum possible value for <i>n</i> index
* @param roaPow powers of R/a up to degree <i>n</i>
* @param ghMSJ G<sup>j</sup><sub>m,s</sub> and H<sup>j</sup><sub>m,s</sub> polynomials
* @param gammaMNS Γ<sup>m</sup><sub>n,s</sub>(γ) function
* @return Components of U<sub>n</sub> derivatives for fixed j, m, s
* @throws OrekitException if some error occurred
*/
private double[][] computeNSum(
final AbsoluteDate date,
final int j,
final int m,
final int s,
final int maxN,
final double[] roaPow,
final GHmsjPolynomials ghMSJ,
final GammaMnsFunction gammaMNS)
throws OrekitException {
// spherical harmonics
final UnnormalizedSphericalHarmonics harmonics = provider.onDate(date);
// Potential derivatives components
double dUdaCos = 0.;
double dUdaSin = 0.;
double dUdhCos = 0.;
double dUdhSin = 0.;
double dUdkCos = 0.;
double dUdkSin = 0.;
double dUdlCos = 0.;
double dUdlSin = 0.;
double dUdAlCos = 0.;
double dUdAlSin = 0.;
double dUdBeCos = 0.;
double dUdBeSin = 0.;
double dUdGaCos = 0.;
double dUdGaSin = 0.;
// I^m
final int Im = I > 0 ? 1 : (m % 2 == 0 ? 1 : -1);
// jacobi v, w, indices from 2.7.1-(15)
final int v = FastMath.abs(m - s);
final int w = FastMath.abs(m + s);
// Initialise lower degree nmin = (Max(2, m, |s|)) for summation over n
final int nmin = FastMath.max(FastMath.max(2, m), FastMath.abs(s));
// Get the corresponding Hansen object
final int sIndex = maxDegree + (j < 0 ? -s : s);
final int jIndex = FastMath.abs(j);
final HansenTesseralLinear hans = this.hansenObjects[sIndex][jIndex];
// n-SUM from nmin to N
for (int n = nmin; n <= maxN; n++) {
// If (n - s) is odd, the contribution is null because of Vmns
if ((n - s) % 2 == 0) {
// Vmns coefficient
final double fns = fact[n + FastMath.abs(s)];
final double vMNS = CoefficientsFactory.getVmns(m, n, s, fns, fact[n - m]);
// Inclination function Gamma and derivative
final double gaMNS = gammaMNS.getValue(m, n, s);
final double dGaMNS = gammaMNS.getDerivative(m, n, s);
// Hansen kernel value and derivative
final double kJNS = hans.getValue(-n - 1, chi);
final double dkJNS = hans.getDerivative(-n - 1, chi);
// Gjms, Hjms polynomials and derivatives
final double gMSJ = ghMSJ.getGmsj(m, s, j);
final double hMSJ = ghMSJ.getHmsj(m, s, j);
final double dGdh = ghMSJ.getdGmsdh(m, s, j);
final double dGdk = ghMSJ.getdGmsdk(m, s, j);
final double dGdA = ghMSJ.getdGmsdAlpha(m, s, j);
final double dGdB = ghMSJ.getdGmsdBeta(m, s, j);
final double dHdh = ghMSJ.getdHmsdh(m, s, j);
final double dHdk = ghMSJ.getdHmsdk(m, s, j);
final double dHdA = ghMSJ.getdHmsdAlpha(m, s, j);
final double dHdB = ghMSJ.getdHmsdBeta(m, s, j);
// Jacobi l-index from 2.7.1-(15)
final int l = FastMath.min(n - m, n - FastMath.abs(s));
// Jacobi polynomial and derivative
final DerivativeStructure jacobi =
JacobiPolynomials.getValue(l, v, w, new DerivativeStructure(1, 1, 0, gamma));
// Geopotential coefficients
final double cnm = harmonics.getUnnormalizedCnm(n, m);
final double snm = harmonics.getUnnormalizedSnm(n, m);
// Common factors from expansion of equations 3.3-4
final double cf_0 = roaPow[n] * Im * vMNS;
final double cf_1 = cf_0 * gaMNS * jacobi.getValue();
final double cf_2 = cf_1 * kJNS;
final double gcPhs = gMSJ * cnm + hMSJ * snm;
final double gsMhc = gMSJ * snm - hMSJ * cnm;
final double dKgcPhsx2 = 2. * dkJNS * gcPhs;
final double dKgsMhcx2 = 2. * dkJNS * gsMhc;
final double dUdaCoef = (n + 1) * cf_2;
final double dUdlCoef = j * cf_2;
final double dUdGaCoef =
cf_0 * kJNS * (jacobi.getValue() * dGaMNS + gaMNS * jacobi.getPartialDerivative(1));
// dU / da components
dUdaCos += dUdaCoef * gcPhs;
dUdaSin += dUdaCoef * gsMhc;
// dU / dh components
dUdhCos += cf_1 * (kJNS * (cnm * dGdh + snm * dHdh) + h * dKgcPhsx2);
dUdhSin += cf_1 * (kJNS * (snm * dGdh - cnm * dHdh) + h * dKgsMhcx2);
// dU / dk components
dUdkCos += cf_1 * (kJNS * (cnm * dGdk + snm * dHdk) + k * dKgcPhsx2);
dUdkSin += cf_1 * (kJNS * (snm * dGdk - cnm * dHdk) + k * dKgsMhcx2);
// dU / dLambda components
dUdlCos += dUdlCoef * gsMhc;
dUdlSin += -dUdlCoef * gcPhs;
// dU / alpha components
dUdAlCos += cf_2 * (dGdA * cnm + dHdA * snm);
dUdAlSin += cf_2 * (dGdA * snm - dHdA * cnm);
// dU / dBeta components
dUdBeCos += cf_2 * (dGdB * cnm + dHdB * snm);
dUdBeSin += cf_2 * (dGdB * snm - dHdB * cnm);
// dU / dGamma components
dUdGaCos += dUdGaCoef * gcPhs;
dUdGaSin += dUdGaCoef * gsMhc;
}
}
return new double[][] {
{dUdaCos, dUdaSin},
{dUdhCos, dUdhSin},
{dUdkCos, dUdkSin},
{dUdlCos, dUdlSin},
{dUdAlCos, dUdAlSin},
{dUdBeCos, dUdBeSin},
{dUdGaCos, dUdGaSin}
};
}
private void doParallelUpload(String bucket, String key, File file, ObjectMetadata metadata)
throws IOException {
long contentLength = file.length();
long numParts = (contentLength + PARALLEL_UPLOAD_SIZE - 1) / PARALLEL_UPLOAD_SIZE;
long partSize = (contentLength / numParts) + 1;
log.info("Uploading {} in {} parts with part size {}", file, numParts, partSize);
final List<PartETag> partETags = Lists.newArrayList();
InitiateMultipartUploadRequest initRequest =
new InitiateMultipartUploadRequest(bucket, key, metadata);
InitiateMultipartUploadResult initResponse = s3Client.initiateMultipartUpload(initRequest);
String uploadID = initResponse.getUploadId();
try {
long filePosition = 0;
int partNumber = 1;
Collection<Future<?>> futures = Lists.newArrayList();
while (filePosition < contentLength) {
final UploadPartRequest uploadRequest =
new UploadPartRequest()
.withBucketName(bucket)
.withKey(key)
.withUploadId(uploadID)
.withPartNumber(partNumber)
.withFileOffset(filePosition)
.withFile(file)
.withPartSize(FastMath.min(partSize, contentLength - filePosition));
futures.add(
executorService.submit(
new Callable<Object>() {
@Override
public Object call() {
String theKey = uploadRequest.getKey();
int thePartNumber = uploadRequest.getPartNumber();
log.info("Starting {} part {}", theKey, thePartNumber);
int failures = 0;
UploadPartResult uploadPartResult;
while (true) {
try {
uploadPartResult = s3Client.uploadPart(uploadRequest);
break;
} catch (AmazonClientException ace) {
if (++failures >= MAX_RETRIES) {
throw ace;
}
log.warn(
"Upload {} part {} failed ({}); retrying",
theKey,
thePartNumber,
ace.getMessage());
}
}
partETags.add(uploadPartResult.getPartETag());
log.info("Finished {} part {}", theKey, thePartNumber);
return null;
}
}));
filePosition += partSize;
partNumber++;
}
for (Future<?> future : futures) {
try {
future.get();
} catch (InterruptedException e) {
throw new IOException(e);
} catch (ExecutionException e) {
throw new IOException(e.getCause());
}
}
CompleteMultipartUploadRequest completeRequest =
new CompleteMultipartUploadRequest(bucket, key, uploadID, partETags);
s3Client.completeMultipartUpload(completeRequest);
} catch (AmazonClientException ace) {
AbortMultipartUploadRequest abortRequest =
new AbortMultipartUploadRequest(bucket, key, uploadID);
s3Client.abortMultipartUpload(abortRequest);
throw ace;
}
}
/**
* {@inheritDoc}
*
* <p>For number of successes {@code m} and sample size {@code n}, the upper bound of the support
* is {@code min(m, n)}.
*
* @return upper bound of the support
*/
public int getSupportUpperBound() {
return FastMath.min(getNumberOfSuccesses(), getSampleSize());
}
/**
* Return the highest domain value for the given hypergeometric distribution parameters.
*
* @param m Number of successes in the population.
* @param k Sample size.
* @return the highest domain value of the hypergeometric distribution.
*/
private int getUpperDomain(int m, int k) {
return FastMath.min(k, m);
}
/**
* Computes the potential U derivatives.
*
* <p>The following elements are computed from expression 3.3 - (4).
*
* <pre>
* dU / da
* dU / dh
* dU / dk
* dU / dλ
* dU / dα
* dU / dβ
* dU / dγ
* </pre>
*
* @param date current date
* @return potential derivatives
* @throws OrekitException if an error occurs
*/
private double[] computeUDerivatives(final AbsoluteDate date) throws OrekitException {
// Potential derivatives
double dUda = 0.;
double dUdh = 0.;
double dUdk = 0.;
double dUdl = 0.;
double dUdAl = 0.;
double dUdBe = 0.;
double dUdGa = 0.;
// Compute only if there is at least one resonant tesseral
if (!resOrders.isEmpty()) {
// Gmsj and Hmsj polynomials
final GHmsjPolynomials ghMSJ = new GHmsjPolynomials(k, h, alpha, beta, I);
// GAMMAmns function
final GammaMnsFunction gammaMNS = new GammaMnsFunction(fact, gamma, I);
// R / a up to power degree
final double[] roaPow = new double[maxDegree + 1];
roaPow[0] = 1.;
for (int i = 1; i <= maxDegree; i++) {
roaPow[i] = roa * roaPow[i - 1];
}
// SUM over resonant terms {j,m}
for (int m : resOrders) {
// Resonant index for the current resonant order
final int j = FastMath.max(1, (int) FastMath.round(ratio * m));
// Phase angle
final double jlMmt = j * lm - m * theta;
final double sinPhi = FastMath.sin(jlMmt);
final double cosPhi = FastMath.cos(jlMmt);
// Potential derivatives components for a given resonant pair {j,m}
double dUdaCos = 0.;
double dUdaSin = 0.;
double dUdhCos = 0.;
double dUdhSin = 0.;
double dUdkCos = 0.;
double dUdkSin = 0.;
double dUdlCos = 0.;
double dUdlSin = 0.;
double dUdAlCos = 0.;
double dUdAlSin = 0.;
double dUdBeCos = 0.;
double dUdBeSin = 0.;
double dUdGaCos = 0.;
double dUdGaSin = 0.;
// s-SUM from -sMin to sMax
final int sMin = FastMath.min(maxEccPow - j, maxDegree);
final int sMax = FastMath.min(maxEccPow + j, maxDegree);
for (int s = 0; s <= sMax; s++) {
// Compute the initial values for Hansen coefficients using newComb operators
this.hansenObjects[s + maxDegree][j].computeInitValues(e2, chi, chi2);
// n-SUM for s positive
final double[][] nSumSpos =
computeNSum(date, j, m, s, maxDegree, roaPow, ghMSJ, gammaMNS);
dUdaCos += nSumSpos[0][0];
dUdaSin += nSumSpos[0][1];
dUdhCos += nSumSpos[1][0];
dUdhSin += nSumSpos[1][1];
dUdkCos += nSumSpos[2][0];
dUdkSin += nSumSpos[2][1];
dUdlCos += nSumSpos[3][0];
dUdlSin += nSumSpos[3][1];
dUdAlCos += nSumSpos[4][0];
dUdAlSin += nSumSpos[4][1];
dUdBeCos += nSumSpos[5][0];
dUdBeSin += nSumSpos[5][1];
dUdGaCos += nSumSpos[6][0];
dUdGaSin += nSumSpos[6][1];
// n-SUM for s negative
if (s > 0 && s <= sMin) {
// Compute the initial values for Hansen coefficients using newComb operators
this.hansenObjects[maxDegree - s][j].computeInitValues(e2, chi, chi2);
final double[][] nSumSneg =
computeNSum(date, j, m, -s, maxDegree, roaPow, ghMSJ, gammaMNS);
dUdaCos += nSumSneg[0][0];
dUdaSin += nSumSneg[0][1];
dUdhCos += nSumSneg[1][0];
dUdhSin += nSumSneg[1][1];
dUdkCos += nSumSneg[2][0];
dUdkSin += nSumSneg[2][1];
dUdlCos += nSumSneg[3][0];
dUdlSin += nSumSneg[3][1];
dUdAlCos += nSumSneg[4][0];
dUdAlSin += nSumSneg[4][1];
dUdBeCos += nSumSneg[5][0];
dUdBeSin += nSumSneg[5][1];
dUdGaCos += nSumSneg[6][0];
dUdGaSin += nSumSneg[6][1];
}
}
// Assembly of potential derivatives componants
dUda += cosPhi * dUdaCos + sinPhi * dUdaSin;
dUdh += cosPhi * dUdhCos + sinPhi * dUdhSin;
dUdk += cosPhi * dUdkCos + sinPhi * dUdkSin;
dUdl += cosPhi * dUdlCos + sinPhi * dUdlSin;
dUdAl += cosPhi * dUdAlCos + sinPhi * dUdAlSin;
dUdBe += cosPhi * dUdBeCos + sinPhi * dUdBeSin;
dUdGa += cosPhi * dUdGaCos + sinPhi * dUdGaSin;
}
dUda *= -moa / a;
dUdh *= moa;
dUdk *= moa;
dUdl *= moa;
dUdAl *= moa;
dUdBe *= moa;
dUdGa *= moa;
}
return new double[] {dUda, dUdh, dUdk, dUdl, dUdAl, dUdBe, dUdGa};
}
/**
* Verifies that nextPoisson(mean) generates an empirical distribution of values consistent with
* PoissonDistributionImpl by generating 1000 values, computing a grouped frequency distribution
* of the observed values and comparing this distribution to the corresponding expected
* distribution computed using PoissonDistributionImpl. Uses ChiSquare test of goodness of fit to
* evaluate the null hypothesis that the distributions are the same. If the null hypothesis can be
* rejected with confidence 1 - alpha, the check fails.
*/
public void checkNextPoissonConsistency(double mean) {
// Generate sample values
final int sampleSize = 1000; // Number of deviates to generate
final int minExpectedCount = 7; // Minimum size of expected bin count
long maxObservedValue = 0;
final double alpha = 0.001; // Probability of false failure
Frequency frequency = new Frequency();
for (int i = 0; i < sampleSize; i++) {
long value = randomData.nextPoisson(mean);
if (value > maxObservedValue) {
maxObservedValue = value;
}
frequency.addValue(value);
}
/*
* Set up bins for chi-square test.
* Ensure expected counts are all at least minExpectedCount.
* Start with upper and lower tail bins.
* Lower bin = [0, lower); Upper bin = [upper, +inf).
*/
PoissonDistribution poissonDistribution = new PoissonDistribution(mean);
int lower = 1;
while (poissonDistribution.cumulativeProbability(lower - 1) * sampleSize < minExpectedCount) {
lower++;
}
int upper = (int) (5 * mean); // Even for mean = 1, not much mass beyond 5
while ((1 - poissonDistribution.cumulativeProbability(upper - 1)) * sampleSize
< minExpectedCount) {
upper--;
}
// Set bin width for interior bins. For poisson, only need to look at end bins.
int binWidth = 0;
boolean widthSufficient = false;
double lowerBinMass = 0;
double upperBinMass = 0;
while (!widthSufficient) {
binWidth++;
lowerBinMass = poissonDistribution.cumulativeProbability(lower - 1, lower + binWidth - 1);
upperBinMass = poissonDistribution.cumulativeProbability(upper - binWidth - 1, upper - 1);
widthSufficient = FastMath.min(lowerBinMass, upperBinMass) * sampleSize >= minExpectedCount;
}
/*
* Determine interior bin bounds. Bins are
* [1, lower = binBounds[0]), [lower, binBounds[1]), [binBounds[1], binBounds[2]), ... ,
* [binBounds[binCount - 2], upper = binBounds[binCount - 1]), [upper, +inf)
*
*/
List<Integer> binBounds = new ArrayList<Integer>();
binBounds.add(lower);
int bound = lower + binWidth;
while (bound < upper - binWidth) {
binBounds.add(bound);
bound += binWidth;
}
binBounds.add(
upper); // The size of bin [binBounds[binCount - 2], upper) satisfies binWidth <= size <
// 2*binWidth.
// Compute observed and expected bin counts
final int binCount = binBounds.size() + 1;
long[] observed = new long[binCount];
double[] expected = new double[binCount];
// Bottom bin
observed[0] = 0;
for (int i = 0; i < lower; i++) {
observed[0] += frequency.getCount(i);
}
expected[0] = poissonDistribution.cumulativeProbability(lower - 1) * sampleSize;
// Top bin
observed[binCount - 1] = 0;
for (int i = upper; i <= maxObservedValue; i++) {
observed[binCount - 1] += frequency.getCount(i);
}
expected[binCount - 1] =
(1 - poissonDistribution.cumulativeProbability(upper - 1)) * sampleSize;
// Interior bins
for (int i = 1; i < binCount - 1; i++) {
observed[i] = 0;
for (int j = binBounds.get(i - 1); j < binBounds.get(i); j++) {
observed[i] += frequency.getCount(j);
} // Expected count is (mass in [binBounds[i-1], binBounds[i])) * sampleSize
expected[i] =
(poissonDistribution.cumulativeProbability(binBounds.get(i) - 1)
- poissonDistribution.cumulativeProbability(binBounds.get(i - 1) - 1))
* sampleSize;
}
// Use chisquare test to verify that generated values are poisson(mean)-distributed
ChiSquareTest chiSquareTest = new ChiSquareTest();
// Fail if we can reject null hypothesis that distributions are the same
if (chiSquareTest.chiSquareTest(expected, observed, alpha)) {
StringBuilder msgBuffer = new StringBuilder();
DecimalFormat df = new DecimalFormat("#.##");
msgBuffer.append("Chisquare test failed for mean = ");
msgBuffer.append(mean);
msgBuffer.append(" p-value = ");
msgBuffer.append(chiSquareTest.chiSquareTest(expected, observed));
msgBuffer.append(" chisquare statistic = ");
msgBuffer.append(chiSquareTest.chiSquare(expected, observed));
msgBuffer.append(". \n");
msgBuffer.append("bin\t\texpected\tobserved\n");
for (int i = 0; i < expected.length; i++) {
msgBuffer.append("[");
msgBuffer.append(i == 0 ? 1 : binBounds.get(i - 1));
msgBuffer.append(",");
msgBuffer.append(i == binBounds.size() ? "inf" : binBounds.get(i));
msgBuffer.append(")");
msgBuffer.append("\t\t");
msgBuffer.append(df.format(expected[i]));
msgBuffer.append("\t\t");
msgBuffer.append(observed[i]);
msgBuffer.append("\n");
}
msgBuffer.append("This test can fail randomly due to sampling error with probability ");
msgBuffer.append(alpha);
msgBuffer.append(".");
Assert.fail(msgBuffer.toString());
}
}
/**
* Computes the Kendall's Tau rank correlation coefficient between the two arrays.
*
* @param xArray first data array
* @param yArray second data array
* @return Returns Kendall's Tau rank correlation coefficient for the two arrays
* @throws DimensionMismatchException if the arrays lengths do not match
*/
public double correlation(final double[] xArray, final double[] yArray)
throws DimensionMismatchException {
if (xArray.length != yArray.length) {
throw new DimensionMismatchException(xArray.length, yArray.length);
}
final int n = xArray.length;
final long numPairs = n * (n - 1l) / 2l;
@SuppressWarnings("unchecked")
Pair<Double, Double>[] pairs = new Pair[n];
for (int i = 0; i < n; i++) {
pairs[i] = new Pair<Double, Double>(xArray[i], yArray[i]);
}
Arrays.sort(
pairs,
new Comparator<Pair<Double, Double>>() {
public int compare(Pair<Double, Double> pair1, Pair<Double, Double> pair2) {
int compareFirst = pair1.getFirst().compareTo(pair2.getFirst());
return compareFirst != 0
? compareFirst
: pair1.getSecond().compareTo(pair2.getSecond());
}
});
int tiedXPairs = 0;
int tiedXYPairs = 0;
int consecutiveXTies = 1;
int consecutiveXYTies = 1;
Pair<Double, Double> prev = pairs[0];
for (int i = 1; i < n; i++) {
final Pair<Double, Double> curr = pairs[i];
if (curr.getFirst().equals(prev.getFirst())) {
consecutiveXTies++;
if (curr.getSecond().equals(prev.getSecond())) {
consecutiveXYTies++;
} else {
tiedXYPairs += consecutiveXYTies * (consecutiveXYTies - 1) / 2;
consecutiveXYTies = 1;
}
} else {
tiedXPairs += consecutiveXTies * (consecutiveXTies - 1) / 2;
consecutiveXTies = 1;
tiedXYPairs += consecutiveXYTies * (consecutiveXYTies - 1) / 2;
consecutiveXYTies = 1;
}
prev = curr;
}
tiedXPairs += consecutiveXTies * (consecutiveXTies - 1) / 2;
tiedXYPairs += consecutiveXYTies * (consecutiveXYTies - 1) / 2;
int swaps = 0;
@SuppressWarnings("unchecked")
Pair<Double, Double>[] pairsDestination = new Pair[n];
for (int segmentSize = 1; segmentSize < n; segmentSize <<= 1) {
for (int offset = 0; offset < n; offset += 2 * segmentSize) {
int i = offset;
final int iEnd = FastMath.min(i + segmentSize, n);
int j = iEnd;
final int jEnd = FastMath.min(j + segmentSize, n);
int copyLocation = offset;
while (i < iEnd || j < jEnd) {
if (i < iEnd) {
if (j < jEnd) {
if (pairs[i].getSecond().compareTo(pairs[j].getSecond()) <= 0) {
pairsDestination[copyLocation] = pairs[i];
i++;
} else {
pairsDestination[copyLocation] = pairs[j];
j++;
swaps += iEnd - i;
}
} else {
pairsDestination[copyLocation] = pairs[i];
i++;
}
} else {
pairsDestination[copyLocation] = pairs[j];
j++;
}
copyLocation++;
}
}
final Pair<Double, Double>[] pairsTemp = pairs;
pairs = pairsDestination;
pairsDestination = pairsTemp;
}
int tiedYPairs = 0;
int consecutiveYTies = 1;
prev = pairs[0];
for (int i = 1; i < n; i++) {
final Pair<Double, Double> curr = pairs[i];
if (curr.getSecond().equals(prev.getSecond())) {
consecutiveYTies++;
} else {
tiedYPairs += consecutiveYTies * (consecutiveYTies - 1) / 2;
consecutiveYTies = 1;
}
prev = curr;
}
tiedYPairs += consecutiveYTies * (consecutiveYTies - 1) / 2;
int concordantMinusDiscordant = numPairs - tiedXPairs - tiedYPairs + tiedXYPairs - 2 * swaps;
return concordantMinusDiscordant
/ FastMath.sqrt((numPairs - tiedXPairs) * (numPairs - tiedYPairs));
}
/** {@inheritDoc} */
public Optimum optimize(final LeastSquaresProblem problem) {
// Pull in relevant data from the problem as locals.
final int nR = problem.getObservationSize(); // Number of observed data.
final int nC = problem.getParameterSize(); // Number of parameters.
// Counters.
final Incrementor iterationCounter = problem.getIterationCounter();
final Incrementor evaluationCounter = problem.getEvaluationCounter();
// Convergence criterion.
final ConvergenceChecker<Evaluation> checker = problem.getConvergenceChecker();
// arrays shared with the other private methods
final int solvedCols = FastMath.min(nR, nC);
/* Parameters evolution direction associated with lmPar. */
double[] lmDir = new double[nC];
/* Levenberg-Marquardt parameter. */
double lmPar = 0;
// local point
double delta = 0;
double xNorm = 0;
double[] diag = new double[nC];
double[] oldX = new double[nC];
double[] oldRes = new double[nR];
double[] qtf = new double[nR];
double[] work1 = new double[nC];
double[] work2 = new double[nC];
double[] work3 = new double[nC];
// Evaluate the function at the starting point and calculate its norm.
evaluationCounter.incrementCount();
// value will be reassigned in the loop
Evaluation current = problem.evaluate(problem.getStart());
double[] currentResiduals = current.getResiduals().toArray();
double currentCost = current.getCost();
double[] currentPoint = current.getPoint().toArray();
// Outer loop.
boolean firstIteration = true;
while (true) {
iterationCounter.incrementCount();
final Evaluation previous = current;
// QR decomposition of the jacobian matrix
final InternalData internalData = qrDecomposition(current.getJacobian(), solvedCols);
final double[][] weightedJacobian = internalData.weightedJacobian;
final int[] permutation = internalData.permutation;
final double[] diagR = internalData.diagR;
final double[] jacNorm = internalData.jacNorm;
// residuals already have weights applied
double[] weightedResidual = currentResiduals;
for (int i = 0; i < nR; i++) {
qtf[i] = weightedResidual[i];
}
// compute Qt.res
qTy(qtf, internalData);
// now we don't need Q anymore,
// so let jacobian contain the R matrix with its diagonal elements
for (int k = 0; k < solvedCols; ++k) {
int pk = permutation[k];
weightedJacobian[k][pk] = diagR[pk];
}
if (firstIteration) {
// scale the point according to the norms of the columns
// of the initial jacobian
xNorm = 0;
for (int k = 0; k < nC; ++k) {
double dk = jacNorm[k];
if (dk == 0) {
dk = 1.0;
}
double xk = dk * currentPoint[k];
xNorm += xk * xk;
diag[k] = dk;
}
xNorm = FastMath.sqrt(xNorm);
// initialize the step bound delta
delta = (xNorm == 0) ? initialStepBoundFactor : (initialStepBoundFactor * xNorm);
}
// check orthogonality between function vector and jacobian columns
double maxCosine = 0;
if (currentCost != 0) {
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
double s = jacNorm[pj];
if (s != 0) {
double sum = 0;
for (int i = 0; i <= j; ++i) {
sum += weightedJacobian[i][pj] * qtf[i];
}
maxCosine = FastMath.max(maxCosine, FastMath.abs(sum) / (s * currentCost));
}
}
}
if (maxCosine <= orthoTolerance) {
// Convergence has been reached.
return new OptimumImpl(current, evaluationCounter.getCount(), iterationCounter.getCount());
}
// rescale if necessary
for (int j = 0; j < nC; ++j) {
diag[j] = FastMath.max(diag[j], jacNorm[j]);
}
// Inner loop.
for (double ratio = 0; ratio < 1.0e-4; ) {
// save the state
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
oldX[pj] = currentPoint[pj];
}
final double previousCost = currentCost;
double[] tmpVec = weightedResidual;
weightedResidual = oldRes;
oldRes = tmpVec;
// determine the Levenberg-Marquardt parameter
lmPar =
determineLMParameter(
qtf, delta, diag, internalData, solvedCols, work1, work2, work3, lmDir, lmPar);
// compute the new point and the norm of the evolution direction
double lmNorm = 0;
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
lmDir[pj] = -lmDir[pj];
currentPoint[pj] = oldX[pj] + lmDir[pj];
double s = diag[pj] * lmDir[pj];
lmNorm += s * s;
}
lmNorm = FastMath.sqrt(lmNorm);
// on the first iteration, adjust the initial step bound.
if (firstIteration) {
delta = FastMath.min(delta, lmNorm);
}
// Evaluate the function at x + p and calculate its norm.
evaluationCounter.incrementCount();
current = problem.evaluate(new ArrayRealVector(currentPoint));
currentResiduals = current.getResiduals().toArray();
currentCost = current.getCost();
currentPoint = current.getPoint().toArray();
// compute the scaled actual reduction
double actRed = -1.0;
if (0.1 * currentCost < previousCost) {
double r = currentCost / previousCost;
actRed = 1.0 - r * r;
}
// compute the scaled predicted reduction
// and the scaled directional derivative
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
double dirJ = lmDir[pj];
work1[j] = 0;
for (int i = 0; i <= j; ++i) {
work1[i] += weightedJacobian[i][pj] * dirJ;
}
}
double coeff1 = 0;
for (int j = 0; j < solvedCols; ++j) {
coeff1 += work1[j] * work1[j];
}
double pc2 = previousCost * previousCost;
coeff1 /= pc2;
double coeff2 = lmPar * lmNorm * lmNorm / pc2;
double preRed = coeff1 + 2 * coeff2;
double dirDer = -(coeff1 + coeff2);
// ratio of the actual to the predicted reduction
ratio = (preRed == 0) ? 0 : (actRed / preRed);
// update the step bound
if (ratio <= 0.25) {
double tmp = (actRed < 0) ? (0.5 * dirDer / (dirDer + 0.5 * actRed)) : 0.5;
if ((0.1 * currentCost >= previousCost) || (tmp < 0.1)) {
tmp = 0.1;
}
delta = tmp * FastMath.min(delta, 10.0 * lmNorm);
lmPar /= tmp;
} else if ((lmPar == 0) || (ratio >= 0.75)) {
delta = 2 * lmNorm;
lmPar *= 0.5;
}
// test for successful iteration.
if (ratio >= 1.0e-4) {
// successful iteration, update the norm
firstIteration = false;
xNorm = 0;
for (int k = 0; k < nC; ++k) {
double xK = diag[k] * currentPoint[k];
xNorm += xK * xK;
}
xNorm = FastMath.sqrt(xNorm);
// tests for convergence.
if (checker != null
&& checker.converged(iterationCounter.getCount(), previous, current)) {
return new OptimumImpl(
current, evaluationCounter.getCount(), iterationCounter.getCount());
}
} else {
// failed iteration, reset the previous values
currentCost = previousCost;
for (int j = 0; j < solvedCols; ++j) {
int pj = permutation[j];
currentPoint[pj] = oldX[pj];
}
tmpVec = weightedResidual;
weightedResidual = oldRes;
oldRes = tmpVec;
// Reset "current" to previous values.
current = previous;
}
// Default convergence criteria.
if ((FastMath.abs(actRed) <= costRelativeTolerance
&& preRed <= costRelativeTolerance
&& ratio <= 2.0)
|| delta <= parRelativeTolerance * xNorm) {
return new OptimumImpl(
current, evaluationCounter.getCount(), iterationCounter.getCount());
}
// tests for termination and stringent tolerances
if (FastMath.abs(actRed) <= TWO_EPS && preRed <= TWO_EPS && ratio <= 2.0) {
throw new ConvergenceException(
LocalizedFormats.TOO_SMALL_COST_RELATIVE_TOLERANCE, costRelativeTolerance);
} else if (delta <= TWO_EPS * xNorm) {
throw new ConvergenceException(
LocalizedFormats.TOO_SMALL_PARAMETERS_RELATIVE_TOLERANCE, parRelativeTolerance);
} else if (maxCosine <= TWO_EPS) {
throw new ConvergenceException(
LocalizedFormats.TOO_SMALL_ORTHOGONALITY_TOLERANCE, orthoTolerance);
}
}
}
}
