// In perspective mode, the warping functions are:
// x' = (a0 + a1 x + a2 y) / (c0 x + c1 y + 1)
// y' = (b0 + b1 x + b2 y) / (c0 x + c1 y + 1)
//
// The following calculates the factors a#, b# and c#.
// We do this by creating a set of eight equations with a#, b# and c# as unknowns.
// The equations are derived by:
// 1. substituting the srcPoints for (x, y);
// 2. substituting the corresponding destPoints for (x', y');
// 3. solving the resulting set of equations, with the factors as unknowns.
//
// The equations are like these:
// a0 x a1 y a2 0 0 0 -xx'c0 -yx'c1 = x'
// 0 0 0 b0 x b1 y b2 -xy'c0 -yy'c1 = y'
// The known factors of left hand side ar put in the 8x8 matrix mxLeft for
// all four point pairs, and the right hand side in the one column matrix mxRight.
// After solving, m_mxWarpFactors contains a0, a1, a2, b0, b1, b2, c0, c1.
private void PreCalc(PointD[] destPoints, PointD[] srcPoints)
{
var mxLeft = new GeneralMatrix(8, 8); //mxLeft.Null();
var mxRight = new GeneralMatrix(8, 1);
var row = 0;
for (int i = 0; i < 4; i++)
{
mxLeft.Array[row][0] = 1.0;
mxLeft.Array[row][1] = srcPoints[i].X;
mxLeft.Array[row][2] = srcPoints[i].Y;
mxLeft.Array[row][6] = - srcPoints[i].X * destPoints[i].X;
mxLeft.Array[row][7] = - srcPoints[i].Y * destPoints[i].X;
mxRight.Array[row][0] = destPoints[i].X;
row++;
mxLeft.Array[row][3] = 1.0f;
mxLeft.Array[row][4] = srcPoints[i].X;
mxLeft.Array[row][5] = srcPoints[i].Y;
mxLeft.Array[row][6] = - srcPoints[i].X * destPoints[i].Y;
mxLeft.Array[row][7] = - srcPoints[i].Y * destPoints[i].Y;
mxRight.Array[row][0] = destPoints[i].Y;
row++;
}
_mxWarpFactors = mxLeft.Solve(mxRight);
}
public void ArrayMultiplyEquals()
{
A = R.Copy();
GeneralMatrix B = GeneralMatrix.Random(A.RowDimension, A.ColumnDimension);
A.ArrayMultiplyEquals(B);
Assert.IsTrue(GeneralTests.Check(A.ArrayRightDivideEquals(B), R));
}
/**
* Computes the Moore–Penrose pseudoinverse using the SVD method.
*
* Modified version of the original implementation by Kim van der Linde.
*/
public static GeneralMatrix pinv(GeneralMatrix x)
{
if (x.Rank() < 1)
return null;
if (x.ColumnDimension > x.RowDimension)
return pinv(x.Transpose()).Transpose();
SingularValueDecomposition svdX = new SingularValueDecomposition(x);
double[] singularValues = svdX.SingularValues;
double tol = Math.Max(x.ColumnDimension, x.RowDimension)
* singularValues[0] * 2E-16;
double[] singularValueReciprocals = new double[singularValues.Count()];
for (int i = 0; i < singularValues.Count(); i++)
singularValueReciprocals[i] = Math.Abs(singularValues[i]) < tol ? 0
: (1.0 / singularValues[i]);
double[][] u = svdX.GetU().Array;
double[][] v = svdX.GetV().Array;
int min = Math.Min(x.ColumnDimension, u[0].Count());
double[][] inverse = new double[x.ColumnDimension][];
for (int i = 0; i < x.ColumnDimension; i++) {
inverse[i] = new double[x.RowDimension];
for (int j = 0; j < u.Count(); j++)
for (int k = 0; k < min; k++)
inverse[i][j] += v[i][k] * singularValueReciprocals[k] * u[j][k];
}
return new GeneralMatrix(inverse);
}
public void FindMinimum()
{
int i;
_xCurrent = FirstStep();
double [] xPrev = new double[_nDim];
for (i=0;i<_nDim;i++)
xPrev[i] = _initial[i];
double [] xNext;
GeneralMatrix hPrev = GeneralMatrix.Identity(_nDim,_nDim);
double currEps=1e5;
_curIter=0;
while (currEps>_epsilon && _curIter<_itMax)
{
_hessian = CalculateNextHessianApproximation(hPrev,xPrev,_xCurrent,_f.GetGradient(xPrev),_f.GetGradient(_xCurrent));
xNext = CalculateNextPoint(_xCurrent,_f.GetGradient(_xCurrent),_hessian);
for (i=0;i<_nDim; i++)
{
xPrev[i] = _xCurrent[i];
_xCurrent[i] = xNext[i];
}
currEps = Diff(_xCurrent,xPrev);
_curIter++;
}
}
public void InitData()
{
A = new GeneralMatrix(columnwise, validld);
R = GeneralMatrix.Random(A.RowDimension, A.ColumnDimension);
S = new GeneralMatrix(columnwise, nonconformld);
O = new GeneralMatrix(A.RowDimension, A.ColumnDimension, 1.0);
}
public GyroCal(AHRS sensor)
{
int i = 0;
InitializeComponent();
this.sensor = sensor;
// Add DataReceived event handler.
sensor.DataReceivedEvent += new DataReceivedDelegate(DataReceivedEventHandler);
sensor.confReceivedEvent += new confReceivedDelegate(confReceivedEventHandler);
confReceivedEventHandler(i);
data_collection_enabled = false;
next_data_index = 0;
loggedData = new double[SAMPLES, 3];
threshold = 1.5 * 3.14159 / 180;
bias = new double[3];
calMat = new double[3, 3];
calMat[0, 0] = 1.0;
calMat[1, 1] = 1.0;
calMat[2, 2] = 1.0;
D = new GeneralMatrix(SAMPLES, 10);
for (i = 0; i < SAMPLES; i++)
{
loggedData[i,0] = 0;
loggedData[i, 1] = 0;
loggedData[i, 2] = 0;
}
}
protected override GeneralMatrix CalculateNextHessianApproximation(GeneralMatrix previousH,
double[]prevX, double[]curX, double[]prevGrad, double[]curGrad)
{
GeneralMatrix currentH = new GeneralMatrix(_nDim,_nDim);
GeneralMatrix cX = new GeneralMatrix(curX,_nDim);
GeneralMatrix pX = new GeneralMatrix(prevX,_nDim);
GeneralMatrix cG = new GeneralMatrix(curGrad,_nDim);
GeneralMatrix pG = new GeneralMatrix(prevGrad,_nDim);
GeneralMatrix dX = cX.Subtract(pX);
GeneralMatrix dG = cG.Subtract(pG);
double aK1 = 1/(dX.Transpose().Multiply(dG).GetElement(0,0));
GeneralMatrix aK2 = dX.Multiply(dX.Transpose());
GeneralMatrix aK = aK2.Multiply(aK1);
double bK1 = -1/(dG.Transpose().Multiply(previousH).Multiply(dG).GetElement(0,0));
GeneralMatrix bK2 = previousH.Multiply(dG).Multiply(dG.Transpose()).Multiply(previousH.Transpose());
GeneralMatrix bK =bK2.Multiply(bK1);
currentH = previousH.Add(aK).Add(bK);
return currentH;
}
public GeneralMatrix CalculateHessian(double[]x)
{
GeneralMatrix hessian = new GeneralMatrix(Dimension,Dimension);
for (int i=0; i<Dimension; i++)
for (int j=0; j<Dimension; j++)
hessian.SetElement(i,j,GetPartialDerivativeVal(i,j,x));
return hessian;
}
public void CholeskyDecomposition1()
{
double[][] pvals = {new double[]{1.0, 1.0, 1.0}, new double[]{1.0, 2.0, 3.0}, new double[]{1.0, 3.0, 6.0}};
GeneralMatrix A = new GeneralMatrix(pvals);
CholeskyDecomposition chol = A.chol();
GeneralMatrix L = chol.GetL();
Assert.IsTrue(GeneralTests.Check(A, L.Multiply(L.Transpose())));
}
public void CholeskyDecomposition2()
{
double[][] pvals = {new double[]{1.0, 1.0, 1.0}, new double[]{1.0, 2.0, 3.0}, new double[]{1.0, 3.0, 6.0}};
GeneralMatrix A = new GeneralMatrix(pvals);
CholeskyDecomposition chol = A.chol();
GeneralMatrix X = chol.Solve(GeneralMatrix.Identity(3, 3));
Assert.IsTrue(GeneralTests.Check(A.Multiply(X), GeneralMatrix.Identity(3, 3)));
}
public void EigenValueDecomposition1()
{
double[][] pvals = {new double[]{1.0, 1.0, 1.0}, new double[]{1.0, 2.0, 3.0}, new double[]{1.0, 3.0, 6.0}};
GeneralMatrix A = new GeneralMatrix(pvals);
EigenvalueDecomposition Eig = A.Eigen();
GeneralMatrix D = Eig.D;
GeneralMatrix V = Eig.GetV();
Assert.IsTrue(GeneralTests.Check(A.Multiply(V), V.Multiply(D)));
}
public void Negative_BadColumnIndexSetGoodRowIndexSet()
{
double[][] avals = {new double[]{1.0, 4.0, 7.0, 10.0}, new double[]{2.0, 5.0, 8.0, 11.0}, new double[]{3.0, 6.0, 9.0, 12.0}};
int[] rowindexset = new int[]{1, 2};
int[] badcolumnindexset = new int[]{1, 2, 4};
GeneralMatrix B = new GeneralMatrix(avals);
M = B.GetMatrix(badrowindexset, columnindexset);
}
public void Negative_BadColumnIndexSet2()
{
int ib = 1, ie = 2; /* index ranges for sub GeneralMatrix */
double[][] avals = {new double[]{1.0, 4.0, 7.0, 10.0}, new double[]{2.0, 5.0, 8.0, 11.0}, new double[]{3.0, 6.0, 9.0, 12.0}};
int[] badrowindexset = new int[]{1, 3};
GeneralMatrix B = new GeneralMatrix(avals);
M = B.GetMatrix(ib, ie + B.RowDimension + 1, columnindexset);
}
//display specified matrix
static void displayMatrix(GeneralMatrix displayMat)
{
for (int i = 0; i < displayMat.RowDimension; i++)
{
for (int j = 0; j < displayMat.ColumnDimension; j++)
{
Console.Write(displayMat.GetElement(i, j) + ",");
}
Console.WriteLine();
}
}
public OneDWrapper(IGradientFunction func, GeneralMatrix pX, GeneralMatrix aX)
{
if (pX==null || aX==null)
throw new ArgumentException("Previous x and alfaX may not be null");
_problemDimension = pX.RowDimension;
_function = func;
_previousX = new GeneralMatrix(pX.ArrayCopy);
_alfaX = new GeneralMatrix(aX.ArrayCopy);
}
/// <summary>
/// Constructor where caller determines the value of all fields (used internally by RoundEigenMetric())
/// </summary>
protected Metric(DotNetMatrix.GeneralMatrix matrix,
DotNetMatrix.EigenvalueDecomposition eig, double[] eigenMetric,
DotNetMatrix.GeneralMatrix invEigMatrix, bool isDiagonal,
bool isEuclidean, bool isAntiEuclidean)
{
m_eigenMetric = eigenMetric;
m_matrix = matrix;
m_eig = eig;
m_eigenMetric = eigenMetric;
m_invEigMatrix = invEigMatrix;
m_isDiagonal = isDiagonal;
m_isEuclidean = isEuclidean;
m_isAntiEuclidean = isAntiEuclidean;
}
public void Hessian()
{
double eps = 1e-5;
double[] vector = {0,0,0};
double[][] result = {new double[]{10.0, 2.0, -2.0},
new double[]{ 2.0, 4.0, 2.0},
new double[]{-2.0, 2.0, 4.0}};
GeneralMatrix expectedMatrix = new GeneralMatrix(result);
TestFunction f = new TestFunction();
GeneralMatrix hessian = f.CalculateHessian(vector);
for (int i=0;i<3; i++)
for (int j=0;j<3; j++)
Assert.IsTrue(System.Math.Abs(hessian.GetElement(i,j)-expectedMatrix.GetElement(i,j))<eps);
}
/// <summary>QR Decomposition, computed by Householder reflections.</summary>
/// <param name="A"> Rectangular matrix
/// </param>
/// <returns> Structure to access R and the Householder vectors and compute Q.
/// </returns>
public QRDecomposition(GeneralMatrix A)
{
// Initialize.
QR = A.ArrayCopy;
m = A.RowDimension;
n = A.ColumnDimension;
Rdiag = new double[n];
// Main loop.
for (int k = 0; k < n; k++)
{
// Compute 2-norm of k-th column without under/overflow.
double nrm = 0;
for (int i = k; i < m; i++)
{
nrm = Maths.Hypot(nrm, QR[i][k]);
}
if (nrm != 0.0)
{
// Form k-th Householder vector.
if (QR[k][k] < 0)
{
nrm = - nrm;
}
for (int i = k; i < m; i++)
{
QR[i][k] /= nrm;
}
QR[k][k] += 1.0;
// Apply transformation to remaining columns.
for (int j = k + 1; j < n; j++)
{
double s = 0.0;
for (int i = k; i < m; i++)
{
s += QR[i][k] * QR[i][j];
}
s = (- s) / QR[k][k];
for (int i = k; i < m; i++)
{
QR[i][j] += s * QR[i][k];
}
}
}
Rdiag[k] = - nrm;
}
}
public Optimizer(int dim, double[]initialPar, IGradientFunction f, double step, double epsilon, int itMax)
{
_f = f;
_epsilon = epsilon;
_itMax = itMax;
_nDim = dim;
_xCurrent = new double[_nDim];
_initial = new double[_nDim];
_step = step;
//initiate the first value of alpha to some random value less than 100;
Random rnd = new Random();
_alpha = rnd.NextDouble()*100;
_hessian = new GeneralMatrix(_nDim,_nDim);
for (int i=0; i<_nDim; i++)
_initial[i] = initialPar[i];
}
/// <summary>Initializes this Metric object from metric matrix (called by constructor) </summary>
/// <param name="m">The NxN metric matrix</param>
private void init(DotNetMatrix.GeneralMatrix m)
{
if (!Util.IsSymmetric(m))
{
throw new Exception("The metric matrix must be symmetric");
}
m_matrix = m.Copy();
// System.Console.WriteLine("m_matrix: " + Util.ToString(m_matrix));
// compute eigen value decomposition
m_eig = new DotNetMatrix.EigenvalueDecomposition(m_matrix);
// System.Console.WriteLine("m_eig: " + Util.ToString(m_eig.GetV()));
m_invEigMatrix = m_eig.GetV().Transpose();
m_eigenMetric = m_eig.RealEigenvalues;
// {
// DotNetMatrix.GeneralMatrix D = Util.Diagonal(m_eigenMetric);
// DotNetMatrix.GeneralMatrix tmp = m_eig.GetV().Multiply(D).Multiply(m_invEigMatrix);
// System.Console.WriteLine("input_matrix = " + Util.ToString(tmp));
// }
m_isDiagonal = Util.IsDiagonal(m_matrix);
if (!m_isDiagonal)
{
m_isEuclidean = m_isAntiEuclidean = false;
}
else
{
m_isEuclidean = m_isAntiEuclidean = true;
for (int i = 0; i < m.RowDimension; i++)
{
if (m_matrix.GetElement(i, i) != 1.0)
{
m_isEuclidean = false;
}
if (m_matrix.GetElement(i, i) != -1.0)
{
m_isAntiEuclidean = false;
}
}
}
}
// In bilinear mode, the warping functions are:
// x' = a0 + a1 x y + a2 x + a3 y
// y' = b0 + b1 x y + b2 x + b3 y
//
// Here, we have two sets of four equations. In the first set, the a# factors
// are the unknowns, in the second set the b# factors.
// The equations are of the form:
// a0 + xy a1 + x a2 + y a3 = x'
// The left hand side is identical for both sets. The right hand side differs.
// Therefore, we can solve them in one operation.
// The left hand side factors are put in the 4x4 matrix mxLeft, the right side
// factors are put in the 4x2 matrix mxRight.
// After solving, the first column of m_mxWarpFactors contains a0, a1, a2, a3; the
// second columne contains b0, b1, b2, b3.
private void PreCalc(PointD[] destPoints, PointD[] srcPoints)
{
var mxLeft = new GeneralMatrix(4, 4);
var mxRight = new GeneralMatrix(4, 2);
for (int row = 0; row < 4; row++)
{
mxLeft.Array[row][0] = 1.0;
mxLeft.Array[row][1] = srcPoints[row].X * srcPoints[row].Y;
mxLeft.Array[row][2] = srcPoints[row].X;
mxLeft.Array[row][3] = srcPoints[row].Y;
mxRight.Array[row][0] = destPoints[row].X;
mxRight.Array[row][1] = destPoints[row].Y;
}
_mxWarpFactors = mxLeft.Solve(mxRight);
}
public double[][] decompose(int svCount, double[][] m)
{
GeneralMatrix mMat = new GeneralMatrix(m);
GeneralMatrix resultM = decompose(svCount, mMat);
double[][] resultD = new double[resultM.RowDimension][];
for (int i = 0; i < resultM.RowDimension; i++)
{
resultD[i] = new double[resultM.ColumnDimension];
for (int j = 0; j < resultM.ColumnDimension; j++)
{
resultD[j][i] = resultM.GetElement(i, j);
if ((j + 1) % 4 == 0) // To remove messed up byte
{
resultD[j][i] = 0;
}
}
}
return resultD;
}
//decrypt line of input
static void Decrypt(string input, GeneralMatrix key, GeneralMatrix invKey, Dictionary<char, int> hillDict, StringBuilder decText)
{
string pt = "";
string[] line = input.Trim().Split(' ');
GeneralMatrix inverse = key.Inverse();
for (int i = 0; i < line.Length; i += 3)
{
//3 at a time
double[] temp = new double[3];
temp[0] = Convert.ToInt32(line[i]);
temp[1] = Convert.ToInt32(line[i + 1]);
temp[2] = Convert.ToInt32(line[i + 2]);
//new 3x1 matrix
GeneralMatrix ctMat = new GeneralMatrix(new double[] { temp[0], temp[1], temp[2] }, 3);
//multiply by inverse
GeneralMatrix ptMat = inverse.Multiply(ctMat);
//mod 31 the result
for (int x = 0; x < ptMat.RowDimension; x++)
{
for (int y = 0; y < ptMat.ColumnDimension; y++)
{
var tempElement = Convert.ToInt32(ptMat.GetElement(x, y)) % 31;
while (tempElement < 0)
tempElement = tempElement + 31;
ptMat.SetElement(x, y, tempElement);
}
}
for (int x = 0; x < ptMat.RowDimension; x++)
{
for (int y = 0; y < ptMat.ColumnDimension; y++)
{
pt += hillDict.FirstOrDefault(z => z.Value == ptMat.GetElement(x, y)).Key;
}
}
}
//replace padding with space characters
decText.Replace("?", " ");
decText.AppendLine(pt);
}
/// <summary>
/// Transforms a basis blade to another basis which is represented by the columns of M.
/// </summary>
/// <param name="a">The basis blade to transform according to <paramref name="M"/>.</param>
/// <param name="M">The matrix to use to transform <paramref name="a"/>.</param>
/// <returns>a list of BasisBlade whose sum represents <paramref name="a"/> on the new basis.</returns>
public static ArrayList Transform(BasisBlade a, DotNetMatrix.GeneralMatrix M)
{
ArrayList A = new ArrayList();
A.Add(new BasisBlade(0, a.scale, a.symScale)); // start with just the scalar part;
int dim = M.RowDimension;
// for each 1 bit: convert to list of blades
int i = 0;
uint b = a.bitmap;
while (b != 0)
{
if ((b & 1) != 0)
{
// take column 'i' out of the matrix, wedge it to 'A'
ArrayList tmp = new ArrayList();
for (int j = 0; j < dim; j++)
{
double m = M.GetElement(j, i);
if (m != 0.0)
{
for (int k = 0; k < A.Count; k++)
{
BasisBlade o = BasisBlade.op((BasisBlade)A[k], new BasisBlade((uint)(1 << j), m));
if (o.scale != 0.0)
{
tmp.Add(o);
}
}
}
}
A = tmp;
}
b >>= 1;
i++;
}
return(A);
}
/// <summary>
///Ctor. In the LMA fit N is the number of data points, M is the number of fit parameters.
///Call <code>fit()</code> to start the actual fitting.
/// </summary>
/// <param name="function">The model function to be fitted. Must be able to take M input parameters.</param>
/// <param name="parameters">The initial guess for the fit parameters, length M.</param>
/// <param name="dataPoints">The data points in an array, <code>double[0 = x, 1 = y][point index]</code>. Size must be <code>double[2][N]</code>.</param>
/// <param name="weights">The weights, normally given as: <code>weights[i] = 1 / sigma_i^2</code>.
/// If you have a bad data point, set its weight to zero. If the given array is null,
/// a new array is created with all elements set to 1.</param>
/// <param name="alpha">A Matrix instance. Must be initiated to (M x M) size.
/// In this case we are using the GeneralMatrix type from the open source JAMA library</param>
/// <param name="argDeltaChi2">delta chi square</param>
/// <param name="argMaxIter">maximum number of iterations</param>
public LMA( LMAFunction function, double[] parameters,
double[][] dataPoints, double[] weights,
GeneralMatrix alpha,
double argDeltaChi2, int argMaxIter)
{
if (dataPoints[0].Length != dataPoints[1].Length)
throw new ArgumentException("Data must have the same number of x and y points.");
if (dataPoints.Length != 2)
throw new ArgumentException("Data point array must be 2 x N");
this.function = function;
this.parameters = parameters;
this.dataPoints = dataPoints;
this.weights = CheckWeights(dataPoints[0].Length, weights);
this.incrementedParameters = new double[parameters.Length];
this.alpha = alpha;
this.beta = new double[parameters.Length];
this.da = new double[parameters.Length];
minDeltaChi2 = argDeltaChi2;
maxIterations = argMaxIter;
lambda = Constants.lambda;
}
/// <summary>
/// Calculates priorities using the Saaty method
/// </summary>
/// <param name="argMatrix"> matrix of choices</param>
/// <returns></returns>
public void ComputePriorities(GeneralMatrix argMatrix)
{
double lMax=0.0;
int n = argMatrix.ColumnDimension;
int m = argMatrix.RowDimension;
if (n!= m)
throw new ArgumentException("Matrix must be symmetrical");
_matrix = new GeneralMatrix(n,1);
PCalc(argMatrix,_matrix);
GeneralMatrix product = FCalc(argMatrix,_matrix);
for (int i=0;i<n; i++)
lMax+=product.GetElement(i,0)/n;
_lambdaMax = lMax;
_consistency = (_lambdaMax-n)/(n-1);
_consistencyRatio = _consistency/Constants.randomIndices[n];
}
/// <summary>Cholesky algorithm for symmetric and positive definite matrix.</summary>
/// <param name="Arg"> Square, symmetric matrix.
/// </param>
/// <returns> Structure to access L and isspd flag.
/// </returns>
public CholeskyDecomposition(GeneralMatrix Arg)
{
// Initialize.
double[][] A = Arg.Array;
n = Arg.RowDimension;
L = new double[n][];
for (int i = 0; i < n; i++)
{
L[i] = new double[n];
}
isspd = (Arg.ColumnDimension == n);
// Main loop.
for (int j = 0; j < n; j++)
{
double[] Lrowj = L[j];
double d = 0.0;
for (int k = 0; k < j; k++)
{
double[] Lrowk = L[k];
double s = 0.0;
for (int i = 0; i < k; i++)
{
s += Lrowk[i] * Lrowj[i];
}
Lrowj[k] = s = (A[j][k] - s) / L[k][k];
d = d + s * s;
isspd = isspd & (A[k][j] == A[j][k]);
}
d = A[j][j] - d;
isspd = isspd & (d > 0.0);
L[j][j] = System.Math.Sqrt(System.Math.Max(d, 0.0));
for (int k = j + 1; k < n; k++)
{
L[j][k] = 0.0;
}
}
}
public double[] decompose(int svCount, double[] m)
{
GeneralMatrix mMat = new GeneralMatrix(m, (int)Math.Sqrt(m.Length));
GeneralMatrix[] resultM = decompose(svCount, mMat);
double[] resultD = new double[resultM[0].RowDimension * resultM[0].ColumnDimension + resultM[1].RowDimension * resultM[1].ColumnDimension];
for (int i = 0; i < resultM[0].RowDimension; i++)
{
for (int j = 0; j < resultM[0].ColumnDimension; j++)
{
resultD[i * resultM[0].ColumnDimension + j] = resultM[0].GetElement(i, j);
}
}
for (int i = 0; i < resultM[1].RowDimension; i++)
{
for (int j = 0; j < resultM[1].ColumnDimension; j++)
{
resultD[resultM[0].RowDimension * resultM[0].ColumnDimension
+ i * resultM[1].ColumnDimension + j] = resultM[1].GetElement(i, j);
}
}
return resultD;
}
//encrypt line of input
static void Encrypt(string input, GeneralMatrix key, Dictionary<char, int> hillDict, StringBuilder encText)
{
string ct = "";
string[] pt = input.Trim().Split(' ');
for (int i = 0; i < pt.Length; i += 3)
{
//encrypt 3 letters at a time
double[] temp = new double[3];
temp[0] = Convert.ToInt32(pt[i]);
temp[1] = Convert.ToInt32(pt[i + 1]);
temp[2] = Convert.ToInt32(pt[i + 2]);
//create plain text matrix, transpose and encrypt it
GeneralMatrix ptMat = new GeneralMatrix(new double[] { temp[0], temp[1], temp[2] }, 3);
GeneralMatrix trnasPTMat = ptMat.Transpose();
GeneralMatrix ctMat = key.Multiply(ptMat);
for (int x = 0; x < ctMat.RowDimension; x++)
{
for (int y = 0; y < ctMat.ColumnDimension; y++)
{
var tempElement = Convert.ToInt32(ctMat.GetElement(x, y)) % 31;
ctMat.SetElement(x, y, tempElement);
}
}
for (int x = 0; x < ctMat.RowDimension; x++)
{
for (int y = 0; y < ctMat.ColumnDimension; y++)
{
ct += hillDict.FirstOrDefault(z => z.Value == ctMat.GetElement(x, y)).Key;
}
}
}
//append to string builder
encText.AppendLine(ct);
}
protected override GeneralMatrix CalculateNextHessianApproximation(GeneralMatrix pH,
double[]prevX, double[]curX, double[]prevGrad, double[]curGrad)
{
GeneralMatrix cH = new GeneralMatrix(_nDim,_nDim);
GeneralMatrix cX = new GeneralMatrix(curX,_nDim);
GeneralMatrix pX = new GeneralMatrix(prevX,_nDim);
GeneralMatrix cG = new GeneralMatrix(curGrad,_nDim);
GeneralMatrix pG = new GeneralMatrix(prevGrad,_nDim);
GeneralMatrix sigma = cX.Subtract(pX);
GeneralMatrix gamma = cG.Subtract(pG);
double sigmaTGamma = sigma.Transpose().Multiply(gamma).GetElement(0,0);
GeneralMatrix hGammaSigmaT = pH.Multiply(gamma.Multiply(sigma.Transpose()));
GeneralMatrix sigmaGammaTH = sigma.Multiply(gamma.Transpose().Multiply(pH));
double gammaTHGamma = (gamma.Transpose().Multiply(pH.Multiply(gamma))).GetElement(0,0);
GeneralMatrix sigmaSigmaT = sigma.Multiply(sigma.Transpose());
GeneralMatrix term1 = (hGammaSigmaT.Add(sigmaGammaTH)).Multiply(1/sigmaTGamma);
GeneralMatrix term2 = (sigmaSigmaT.Multiply(1/sigmaTGamma)).Multiply(1+gammaTHGamma/sigmaTGamma);
return pH.Subtract(term1).Add(term2);
}
/// <summary>Solve A*X = B</summary>
/// <param name="B"> right hand side
/// </param>
/// <returns> solution if A is square, least squares solution otherwise
/// </returns>
public virtual GeneralMatrix Solve(GeneralMatrix B)
{
return(m == n ? (new LUDecomposition(this)).Solve(B):(new QRDecomposition(this)).Solve(B));
}
public void DEF()
{
double[][] rankdef = {new double[]{1.0, 4.0, 7.0, 10.0}, new double[]{2.0, 5.0, 8.0, 11.0}, new double[]{3.0, 6.0, 9.0, 12.0}};
GeneralMatrix def = new GeneralMatrix(rankdef);
Assert.IsTrue(GeneralTests.Check(def.Rank(), System.Math.Min(def.RowDimension, def.ColumnDimension) - 1));
}
/// <summary>Check for symmetry, then construct the eigenvalue decomposition</summary>
/// <param name="Arg"> Square matrix
/// </param>
/// <returns> Structure to access D and V.
/// </returns>
public EigenvalueDecomposition(GeneralMatrix Arg)
{
double[][] A = Arg.Array;
n = Arg.ColumnDimension;
V = new double[n][];
for (int i = 0; i < n; i++)
{
V[i] = new double[n];
}
d = new double[n];
e = new double[n];
issymmetric = true;
for (int j = 0; (j < n) & issymmetric; j++)
{
for (int i = 0; (i < n) & issymmetric; i++)
{
issymmetric = (A[i][j] == A[j][i]);
}
}
if (issymmetric)
{
for (int i = 0; i < n; i++)
{
for (int j = 0; j < n; j++)
{
V[i][j] = A[i][j];
}
}
// Tridiagonalize.
tred2();
// Diagonalize.
tql2();
}
else
{
H = new double[n][];
for (int i2 = 0; i2 < n; i2++)
{
H[i2] = new double[n];
}
ort = new double[n];
for (int j = 0; j < n; j++)
{
for (int i = 0; i < n; i++)
{
H[i][j] = A[i][j];
}
}
// Reduce to Hessenberg form.
orthes();
// Reduce Hessenberg to real Schur form.
hqr2();
}
}
public void SolveLinear()
{
double[][] subavals = {new double[]{5.0, 8.0, 11.0}, new double[]{6.0, 9.0, 12.0}};
double[][] sqSolution = {new double[]{13.0}, new double[]{15.0}};
GeneralMatrix sub = new GeneralMatrix(subavals);
GeneralMatrix o = new GeneralMatrix(sub.RowDimension, 1, 1.0);
GeneralMatrix sol = new GeneralMatrix(sqSolution);
GeneralMatrix sq = sub.GetMatrix(0, sub.RowDimension - 1, 0, sub.RowDimension - 1);
Assert.IsTrue(GeneralTests.Check(sq.Solve(sol), o));
}
public void SingularValues()
{
double[][] condmat = {new double[]{1.0, 3.0}, new double[]{7.0, 9.0}};
GeneralMatrix B = new GeneralMatrix(condmat);
SingularValueDecomposition SVD = B.SVD();
double[] singularvalues = SVD.SingularValues;
Assert.IsTrue(GeneralTests.Check(B.Condition(), singularvalues[0] / singularvalues[System.Math.Min(B.RowDimension, B.ColumnDimension) - 1]));
}
/// <summary>Construct the singular value decomposition</summary>
/// <param name="Arg"> Rectangular matrix
/// </param>
/// <returns> Structure to access U, S and V.
/// </returns>
public SingularValueDecomposition(GeneralMatrix Arg)
{
// Derived from LINPACK code.
// Initialize.
double[][] A = Arg.ArrayCopy;
m = Arg.RowDimension;
n = Arg.ColumnDimension;
int nu = System.Math.Min(m, n);
s = new double[System.Math.Min(m + 1, n)];
U = new double[m][];
for (int i = 0; i < m; i++)
{
U[i] = new double[nu];
}
V = new double[n][];
for (int i2 = 0; i2 < n; i2++)
{
V[i2] = new double[n];
}
double[] e = new double[n];
double[] work = new double[m];
bool wantu = true;
bool wantv = true;
// Reduce A to bidiagonal form, storing the diagonal elements
// in s and the super-diagonal elements in e.
int nct = System.Math.Min(m - 1, n);
int nrt = System.Math.Max(0, System.Math.Min(n - 2, m));
for (int k = 0; k < System.Math.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.
s[k] = 0;
for (int i = k; i < m; i++)
{
s[k] = Maths.Hypot(s[k], A[i][k]);
}
if (s[k] != 0.0)
{
if (A[k][k] < 0.0)
{
s[k] = -s[k];
}
for (int i = k; i < m; i++)
{
A[i][k] /= s[k];
}
A[k][k] += 1.0;
}
s[k] = -s[k];
}
for (int j = k + 1; j < n; j++)
{
if ((k < nct) & (s[k] != 0.0))
{
// 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 (wantu & (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] = Maths.Hypot(e[k], e[i]);
}
if (e[k] != 0.0)
{
if (e[k + 1] < 0.0)
{
e[k] = -e[k];
}
for (int i = k + 1; i < n; i++)
{
e[i] /= e[k];
}
e[k + 1] += 1.0;
}
e[k] = -e[k];
if ((k + 1 < m) & (e[k] != 0.0))
{
// Apply the transformation.
for (int i = k + 1; i < m; i++)
{
work[i] = 0.0;
}
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++)
{
double t = (-e[j]) / e[k + 1];
for (int i = k + 1; i < m; i++)
{
A[i][j] += t * work[i];
}
}
}
if (wantv)
{
// 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 = System.Math.Min(n, m + 1);
if (nct < n)
{
s[nct] = A[nct][nct];
}
if (m < p)
{
s[p - 1] = 0.0;
}
if (nrt + 1 < p)
{
e[nrt] = A[nrt][p - 1];
}
e[p - 1] = 0.0;
// If required, generate U.
if (wantu)
{
for (int j = nct; j < nu; j++)
{
for (int i = 0; i < m; i++)
{
U[i][j] = 0.0;
}
U[j][j] = 1.0;
}
for (int k = nct - 1; k >= 0; k--)
{
if (s[k] != 0.0)
{
for (int j = k + 1; j < nu; 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] = 1.0 + U[k][k];
for (int i = 0; i < k - 1; i++)
{
U[i][k] = 0.0;
}
}
else
{
for (int i = 0; i < m; i++)
{
U[i][k] = 0.0;
}
U[k][k] = 1.0;
}
}
}
// If required, generate V.
if (wantv)
{
for (int k = n - 1; k >= 0; k--)
{
if ((k < nrt) & (e[k] != 0.0))
{
for (int j = k + 1; j < nu; 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] = 0.0;
}
V[k][k] = 1.0;
}
}
// Main iteration loop for the singular values.
int pp = p - 1;
int iter = 0;
double eps = System.Math.Pow(2.0, -52.0);
while (p > 0)
{
int k, 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 >= -1; k--)
{
if (k == -1)
{
break;
}
if (System.Math.Abs(e[k]) <= eps * (System.Math.Abs(s[k]) + System.Math.Abs(s[k + 1])))
{
e[k] = 0.0;
break;
}
}
if (k == p - 2)
{
kase = 4;
}
else
{
int ks;
for (ks = p - 1; ks >= k; ks--)
{
if (ks == k)
{
break;
}
double t = (ks != p?System.Math.Abs(e[ks]):0.0) + (ks != k + 1?System.Math.Abs(e[ks - 1]):0.0);
if (System.Math.Abs(s[ks]) <= eps * t)
{
s[ks] = 0.0;
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] = 0.0;
for (int j = p - 2; j >= k; j--)
{
double t = Maths.Hypot(s[j], f);
double cs = s[j] / t;
double sn = f / t;
s[j] = t;
if (j != k)
{
f = (-sn) * e[j - 1];
e[j - 1] = cs * e[j - 1];
}
if (wantv)
{
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] = 0.0;
for (int j = k; j < p; j++)
{
double t = Maths.Hypot(s[j], f);
double cs = s[j] / t;
double sn = f / t;
s[j] = t;
f = (-sn) * e[j];
e[j] = cs * e[j];
if (wantu)
{
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.
double scale = System.Math.Max(System.Math.Max(System.Math.Max(System.Math.Max(System.Math.Abs(s[p - 1]), System.Math.Abs(s[p - 2])), System.Math.Abs(e[p - 2])), System.Math.Abs(s[k])), System.Math.Abs(e[k]));
double sp = s[p - 1] / scale;
double spm1 = s[p - 2] / scale;
double epm1 = e[p - 2] / scale;
double sk = s[k] / scale;
double ek = e[k] / scale;
double b = ((spm1 + sp) * (spm1 - sp) + epm1 * epm1) / 2.0;
double c = (sp * epm1) * (sp * epm1);
double shift = 0.0;
if ((b != 0.0) | (c != 0.0))
{
shift = System.Math.Sqrt(b * b + c);
if (b < 0.0)
{
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 = Maths.Hypot(f, g);
double cs = f / t;
double sn = g / t;
if (j != k)
{
e[j - 1] = t;
}
f = cs * s[j] + sn * e[j];
e[j] = cs * e[j] - sn * s[j];
g = sn * s[j + 1];
s[j + 1] = cs * s[j + 1];
if (wantv)
{
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 = Maths.Hypot(f, g);
cs = f / t;
sn = g / t;
s[j] = t;
f = cs * e[j] + sn * s[j + 1];
s[j + 1] = (-sn) * e[j] + cs * s[j + 1];
g = sn * e[j + 1];
e[j + 1] = cs * e[j + 1];
if (wantu && (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.
case 4:
{
// Make the singular values positive.
if (s[k] <= 0.0)
{
s[k] = (s[k] < 0.0?-s[k]:0.0);
if (wantv)
{
for (int i = 0; i <= pp; i++)
{
V[i][k] = -V[i][k];
}
}
}
// Order the singular values.
while (k < pp)
{
if (s[k] >= s[k + 1])
{
break;
}
double t = s[k];
s[k] = s[k + 1];
s[k + 1] = t;
if (wantv && (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 (wantu && (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;
}
}
}
/// <summary>Solve X*A = B, which is also A'*X' = B'</summary>
/// <param name="B"> right hand side
/// </param>
/// <returns> solution if A is square, least squares solution otherwise.
/// </returns>
public virtual GeneralMatrix SolveTranspose(GeneralMatrix B)
{
return(Transpose().Solve(B.Transpose()));
}
/// <summary>LU Decomposition</summary>
/// <param name="A"> Rectangular matrix
/// </param>
/// <returns> Structure to access L, U and piv.
/// </returns>
public LUDecomposition(GeneralMatrix A)
{
// Use a "left-looking", dot-product, Crout/Doolittle algorithm.
LU = A.ArrayCopy;
m = A.RowDimension;
n = A.ColumnDimension;
piv = new int[m];
for (int i = 0; i < m; i++)
{
piv[i] = i;
}
pivsign = 1;
double[] LUrowi;
double[] LUcolj = new double[m];
// Outer loop.
for (int j = 0; j < n; j++)
{
// Make a copy of the j-th column to localize references.
for (int i = 0; i < m; i++)
{
LUcolj[i] = LU[i][j];
}
// Apply previous transformations.
for (int i = 0; i < m; i++)
{
LUrowi = LU[i];
// Most of the time is spent in the following dot product.
int kmax = System.Math.Min(i, j);
double s = 0.0;
for (int k = 0; k < kmax; k++)
{
s += LUrowi[k] * LUcolj[k];
}
LUrowi[j] = LUcolj[i] -= s;
}
// Find pivot and exchange if necessary.
int p = j;
for (int i = j + 1; i < m; i++)
{
if (System.Math.Abs(LUcolj[i]) > System.Math.Abs(LUcolj[p]))
{
p = i;
}
}
if (p != j)
{
for (int k = 0; k < n; k++)
{
double t = LU[p][k]; LU[p][k] = LU[j][k]; LU[j][k] = t;
}
int k2 = piv[p]; piv[p] = piv[j]; piv[j] = k2;
pivsign = -pivsign;
}
// Compute multipliers.
if (j < m & LU[j][j] != 0.0)
{
for (int i = j + 1; i < m; i++)
{
LU[i][j] /= LU[j][j];
}
}
}
}
protected RpropResult RPropLoop(double[] seed, bool precise)
{
//Console.WriteLine("RpropLoop");
InitialStepSize();
double[] curGradient;
RpropResult ret = new RpropResult();
if (seed != null)
{
curGradient = InitialPointFromSeed(ret, seed);
}
else
{
curGradient = InitialPoint(ret);
}
double curUtil = ret.initialUtil;
double oldUtil = curUtil;
double[] formerGradient = new double[dim];
double[] curValue = new double[dim];
double[] testValue = new double[dim];
double lambda = 0.1;
Tuple <double[], double> tup;
Buffer.BlockCopy(ret.initialValue, 0, curValue, 0, sizeof(double) * dim);
formerGradient = curGradient;
//Buffer.BlockCopy(curGradient,0,formerGradient,0,sizeof(double)*dim);
int itcounter = 0;
int badcounter = 0;
/*Console.WriteLine("Initial Sol:");
* for(int i=0; i<dim;i++) {
* Console.Write("{0} ",curValue[i]);
* }
* Console.WriteLine();
* Console.WriteLine("Initial Util: {0}",curUtil);
*/
#if (GSOLVER_LOG)
Log(curUtil, curValue);
#endif
int maxIter = 60;
int maxBad = 30;
double minStep = 1E-11;
if (precise)
{
maxIter = 120; //110
maxBad = 60; //60
minStep = 1E-15; //15
}
int convergendDims = 0;
while (itcounter++ < maxIter && badcounter < maxBad)
{
convergendDims = 0;
//First Order resp. approximated Second Order Gradient
for (int i = 0; i < dim; i++)
{
if (curGradient[i] * formerGradient[i] > 0)
{
rpropStepWidth[i] *= 1.3;
}
else if (curGradient[i] * formerGradient[i] < 0)
{
rpropStepWidth[i] *= 0.5;
}
rpropStepWidth[i] = Math.Max(minStep, rpropStepWidth[i]);
//rpropStepWidth[i] = Math.Max(0.000001,rpropStepWidth[i]);
if (curUtil > 0.0)
{
//if (curGradient[i] > 0) curValue[i] += rpropStepWidth[i];
//else if (curGradient[i] < 0) curValue[i] -= rpropStepWidth[i];
}
else
{
//linear assumption
//curValue[i] += -curUtil/curGradient[i];
//quadratic assumption
/*double ypSquare = 0;
* for(int j=0; j<dim; j++) {
* ypSquare += curGradient[j]*curGradient[j];
* }
* double m=ypSquare/curUtil;
* curValue[i] = -curGradient[i]/(2*m) + curValue[i];*/
}
if (curValue[i] > limits[i, 1])
{
curValue[i] = limits[i, 1];
}
else if (curValue[i] < limits[i, 0])
{
curValue[i] = limits[i, 0];
}
if (rpropStepWidth[i] < rpropStepConvergenceThreshold[i])
{
++convergendDims;
}
}
//Abort if all dimensions are converged
if (!precise && convergendDims >= dim)
{
if (curUtil > ret.finalUtil)
{
ret.finalUtil = curUtil;
Buffer.BlockCopy(curValue, 0, ret.finalValue, 0, sizeof(double) * dim);
}
return(ret);
}
// Conjugate Gradient
if (curUtil < 0.5)
{
DotNetMatrix.GeneralMatrix X = new DotNetMatrix.GeneralMatrix(dim * 2 + 1, dim);
DotNetMatrix.GeneralMatrix Y = new DotNetMatrix.GeneralMatrix(dim * 2 + 1, 1);
for (int n = 0; n < dim; n++)
{
X.SetElement(0, n, curValue[n]);
}
Y.SetElement(0, 0, 0);
for (int j = 0; j < dim * 2; j++)
{
for (int n = 0; n < dim; n++)
{
double rVal = rand.NextDouble() * rpropStepConvergenceThreshold[n] * 1000.0;
testValue[n] = curValue[n] + rVal;
}
tup = term.Differentiate(testValue);
for (int n = 0; n < dim; n++)
{
X.SetElement(j + 1, n, tup.Item1[n]);
}
Y.SetElement(j + 1, 0, tup.Item2 - curUtil);
}
DotNetMatrix.GeneralMatrix JJ = X.Transpose().Multiply(X);
/*if(curUtil>oldUtil) lambda *= 10;
* else lambda *= 0.1;*/
//DotNetMatrix.GeneralMatrix B = JJ.Add(GeneralMatrix.Identity(dim, dim).Multiply(lambda)).Inverse().Multiply(X.Transpose()).Multiply(Y);
DotNetMatrix.GeneralMatrix B = JJ.Add(JJ.SVD().S.Multiply(lambda)).Inverse().Multiply(X.Transpose()).Multiply(Y);
//DotNetMatrix.GeneralMatrix B = JJ.Inverse().Multiply(X.Transpose()).Multiply(Y);
for (int j = 0; j < dim; j++)
{
curValue[j] += 0.01 * B.GetElement(j, 0);
}
Console.WriteLine(curUtil);
Console.Write(B.Transpose());
Console.WriteLine();
}
/////////////////////////
this.FEvals++;
tup = term.Differentiate(curValue);
bool allZero = true;
for (int i = 0; i < dim; i++)
{
if (Double.IsNaN(tup.Item1[i]))
{
ret.aborted = true;
#if (GSOLVER_LOG)
LogStep();
#endif
return(ret);
}
allZero &= (tup.Item1[i] == 0);
}
oldUtil = curUtil;
curUtil = tup.Item2;
formerGradient = curGradient;
curGradient = tup.Item1;
#if (GSOLVER_LOG)
Log(curUtil, curValue);
#endif
//Console.WriteLine("CurUtil: {0} Final {1}",curUtil,ret.finalUtil);
if (curUtil > ret.finalUtil)
{
badcounter = 0; //Math.Max(0,badcounter-1);
//if (curUtil-ret.finalUtil < 0.00000000000001) {
//Console.WriteLine("not better");
// badcounter++;
//} else {
//badcounter = 0;
//}
ret.finalUtil = curUtil;
Buffer.BlockCopy(curValue, 0, ret.finalValue, 0, sizeof(double) * dim);
//ret.finalValue = curValue;
#if (ALWAYS_CHECK_THRESHOLD)
if (curUtil > utilityThreshold)
{
return(ret);
}
#endif
}
else
{
//if (curUtil < ret.finalUtil || curUtil > 0) badcounter++;
badcounter++;
}
if (allZero)
{
//Console.WriteLine("All Zero!");
/*Console.WriteLine("Util {0}",curUtil);
* Console.Write("Vals: ");
* for(int i=0; i < dim; i++) {
* Console.Write("{0}\t",curValue[i]);
* }
* Console.WriteLine();*/
ret.aborted = false;
#if (GSOLVER_LOG)
LogStep();
#endif
return(ret);
}
}
#if (GSOLVER_LOG)
LogStep();
#endif
ret.aborted = false;
return(ret);
}
/// <summary>Creates a new instance of Metric</summary>
/// <param name="m">The NxN metric matrix</param>
public Metric(DotNetMatrix.GeneralMatrix m)
{
init(m); // forward call to init()
}
