/*************************************************************************
* Unsets HQRNDState structure
*************************************************************************/
private static void unsetstate(ref hqrnd.hqrndstate state)
{
state.s1 = 0;
state.s2 = 0;
state.v = 0;
state.magicv = 0;
}
/*************************************************************************
* Generation of random NxN complex matrix with given condition number C and
* norm2(A)=1
*
* INPUT PARAMETERS:
* N - matrix size
* C - condition number (in 2-norm)
*
* OUTPUT PARAMETERS:
* A - random matrix with norm2(A)=1 and cond(A)=C
*
* -- ALGLIB routine --
* 04.12.2009
* Bochkanov Sergey
*************************************************************************/
public static void cmatrixrndcond(int n,
double c,
ref AP.Complex[,] a)
{
int i = 0;
int j = 0;
double l1 = 0;
double l2 = 0;
hqrnd.hqrndstate state = new hqrnd.hqrndstate();
AP.Complex v = 0;
System.Diagnostics.Debug.Assert(n >= 1 & (double)(c) >= (double)(1), "CMatrixRndCond: N<1 or C<1!");
a = new AP.Complex[n - 1 + 1, n - 1 + 1];
if (n == 1)
{
//
// special case
//
hqrnd.hqrndrandomize(ref state);
hqrnd.hqrndunit2(ref state, ref v.x, ref v.y);
a[0, 0] = v;
return;
}
l1 = 0;
l2 = Math.Log(1 / c);
for (i = 0; i <= n - 1; i++)
{
for (j = 0; j <= n - 1; j++)
{
a[i, j] = 0;
}
}
a[0, 0] = Math.Exp(l1);
for (i = 1; i <= n - 2; i++)
{
a[i, i] = Math.Exp(AP.Math.RandomReal() * (l2 - l1) + l1);
}
a[n - 1, n - 1] = Math.Exp(l2);
cmatrixrndorthogonalfromtheleft(ref a, n, n);
cmatrixrndorthogonalfromtheright(ref a, n, n);
}
public static bool testhqrnd(bool silent)
{
bool result = new bool();
bool waserrors = new bool();
int samplesize = 0;
double sigmathreshold = 0;
int passcount = 0;
int n = 0;
int i = 0;
int pass = 0;
int s1 = 0;
int s2 = 0;
int i1 = 0;
int i2 = 0;
double r1 = 0;
double r2 = 0;
double[] x = new double[0];
double mean = 0;
double means = 0;
double stddev = 0;
double stddevs = 0;
double lambda = 0;
bool seederrors = new bool();
bool urerrors = new bool();
double ursigmaerr = 0;
bool uierrors = new bool();
double uisigmaerr = 0;
bool normerrors = new bool();
double normsigmaerr = 0;
bool experrors = new bool();
double expsigmaerr = 0;
hqrnd.hqrndstate state = new hqrnd.hqrndstate();
waserrors = false;
sigmathreshold = 7;
samplesize = 100000;
passcount = 50;
x = new double[samplesize - 1 + 1];
//
// Test seed errors
//
seederrors = false;
for (pass = 1; pass <= passcount; pass++)
{
s1 = 1 + AP.Math.RandomInteger(32000);
s2 = 1 + AP.Math.RandomInteger(32000);
unsetstate(ref state);
hqrnd.hqrndseed(s1, s2, ref state);
i1 = hqrnd.hqrnduniformi(100, ref state);
unsetstate(ref state);
hqrnd.hqrndseed(s1, s2, ref state);
i2 = hqrnd.hqrnduniformi(100, ref state);
seederrors = seederrors | i1 != i2;
unsetstate(ref state);
hqrnd.hqrndseed(s1, s2, ref state);
r1 = hqrnd.hqrnduniformr(ref state);
unsetstate(ref state);
hqrnd.hqrndseed(s1, s2, ref state);
r2 = hqrnd.hqrnduniformr(ref state);
seederrors = seederrors | (double)(r1) != (double)(r2);
}
//
// Test HQRNDRandomize() and real uniform generator
//
unsetstate(ref state);
hqrnd.hqrndrandomize(ref state);
urerrors = false;
ursigmaerr = 0;
for (i = 0; i <= samplesize - 1; i++)
{
x[i] = hqrnd.hqrnduniformr(ref state);
}
for (i = 0; i <= samplesize - 1; i++)
{
urerrors = urerrors | (double)(x[i]) <= (double)(0) | (double)(x[i]) >= (double)(1);
}
calculatemv(ref x, samplesize, ref mean, ref means, ref stddev, ref stddevs);
if ((double)(means) != (double)(0))
{
ursigmaerr = Math.Max(ursigmaerr, Math.Abs((mean - 0.5) / means));
}
else
{
urerrors = true;
}
if ((double)(stddevs) != (double)(0))
{
ursigmaerr = Math.Max(ursigmaerr, Math.Abs((stddev - Math.Sqrt((double)(1) / (double)(12))) / stddevs));
}
else
{
urerrors = true;
}
urerrors = urerrors | (double)(ursigmaerr) > (double)(sigmathreshold);
//
// Test HQRNDRandomize() and integer uniform
//
unsetstate(ref state);
hqrnd.hqrndrandomize(ref state);
uierrors = false;
uisigmaerr = 0;
for (n = 2; n <= 10; n++)
{
for (i = 0; i <= samplesize - 1; i++)
{
x[i] = hqrnd.hqrnduniformi(n, ref state);
}
for (i = 0; i <= samplesize - 1; i++)
{
uierrors = uierrors | (double)(x[i]) < (double)(0) | (double)(x[i]) >= (double)(n);
}
calculatemv(ref x, samplesize, ref mean, ref means, ref stddev, ref stddevs);
if ((double)(means) != (double)(0))
{
uisigmaerr = Math.Max(uisigmaerr, Math.Abs((mean - 0.5 * (n - 1)) / means));
}
else
{
uierrors = true;
}
if ((double)(stddevs) != (double)(0))
{
uisigmaerr = Math.Max(uisigmaerr, Math.Abs((stddev - Math.Sqrt((AP.Math.Sqr(n) - 1) / 12)) / stddevs));
}
else
{
uierrors = true;
}
}
uierrors = uierrors | (double)(uisigmaerr) > (double)(sigmathreshold);
//
// Special 'close-to-limit' test on uniformity of integers
// (straightforward implementation like 'RND mod N' will return
// non-uniform numbers for N=2/3*LIMIT)
//
unsetstate(ref state);
hqrnd.hqrndrandomize(ref state);
uierrors = false;
uisigmaerr = 0;
n = (int)Math.Round(2.0 / 3.0 * 2147483563.0);
for (i = 0; i <= samplesize - 1; i++)
{
x[i] = hqrnd.hqrnduniformi(n, ref state);
}
for (i = 0; i <= samplesize - 1; i++)
{
uierrors = uierrors | (double)(x[i]) < (double)(0) | (double)(x[i]) >= (double)(n);
}
calculatemv(ref x, samplesize, ref mean, ref means, ref stddev, ref stddevs);
if ((double)(means) != (double)(0))
{
uisigmaerr = Math.Max(uisigmaerr, Math.Abs((mean - 0.5 * (n - 1)) / means));
}
else
{
uierrors = true;
}
if ((double)(stddevs) != (double)(0))
{
uisigmaerr = Math.Max(uisigmaerr, Math.Abs((stddev - Math.Sqrt((AP.Math.Sqr(n) - 1) / 12)) / stddevs));
}
else
{
uierrors = true;
}
uierrors = uierrors | (double)(uisigmaerr) > (double)(sigmathreshold);
//
// Test normal
//
unsetstate(ref state);
hqrnd.hqrndrandomize(ref state);
normerrors = false;
normsigmaerr = 0;
i = 0;
while (i < samplesize)
{
hqrnd.hqrndnormal2(ref state, ref r1, ref r2);
x[i] = r1;
if (i + 1 < samplesize)
{
x[i + 1] = r2;
}
i = i + 2;
}
calculatemv(ref x, samplesize, ref mean, ref means, ref stddev, ref stddevs);
if ((double)(means) != (double)(0))
{
normsigmaerr = Math.Max(normsigmaerr, Math.Abs((mean - 0) / means));
}
else
{
normerrors = true;
}
if ((double)(stddevs) != (double)(0))
{
normsigmaerr = Math.Max(normsigmaerr, Math.Abs((stddev - 1) / stddevs));
}
else
{
normerrors = true;
}
normerrors = normerrors | (double)(normsigmaerr) > (double)(sigmathreshold);
//
// Test exponential
//
unsetstate(ref state);
hqrnd.hqrndrandomize(ref state);
experrors = false;
expsigmaerr = 0;
lambda = 2 + 5 * AP.Math.RandomReal();
for (i = 0; i <= samplesize - 1; i++)
{
x[i] = hqrnd.hqrndexponential(lambda, ref state);
}
for (i = 0; i <= samplesize - 1; i++)
{
uierrors = uierrors | (double)(x[i]) < (double)(0);
}
calculatemv(ref x, samplesize, ref mean, ref means, ref stddev, ref stddevs);
if ((double)(means) != (double)(0))
{
expsigmaerr = Math.Max(expsigmaerr, Math.Abs((mean - 1.0 / lambda) / means));
}
else
{
experrors = true;
}
if ((double)(stddevs) != (double)(0))
{
expsigmaerr = Math.Max(expsigmaerr, Math.Abs((stddev - 1.0 / lambda) / stddevs));
}
else
{
experrors = true;
}
experrors = experrors | (double)(expsigmaerr) > (double)(sigmathreshold);
//
// Final report
//
waserrors = seederrors | urerrors | uierrors | normerrors | experrors;
if (!silent)
{
System.Console.Write("RNG TEST");
System.Console.WriteLine();
System.Console.Write("SEED TEST: ");
if (!seederrors)
{
System.Console.Write("OK");
System.Console.WriteLine();
}
else
{
System.Console.Write("FAILED");
System.Console.WriteLine();
}
System.Console.Write("UNIFORM CONTINUOUS: ");
if (!urerrors)
{
System.Console.Write("OK");
System.Console.WriteLine();
}
else
{
System.Console.Write("FAILED");
System.Console.WriteLine();
}
System.Console.Write("UNIFORM INTEGER: ");
if (!uierrors)
{
System.Console.Write("OK");
System.Console.WriteLine();
}
else
{
System.Console.Write("FAILED");
System.Console.WriteLine();
}
System.Console.Write("NORMAL: ");
if (!normerrors)
{
System.Console.Write("OK");
System.Console.WriteLine();
}
else
{
System.Console.Write("FAILED");
System.Console.WriteLine();
}
System.Console.Write("EXPONENTIAL: ");
if (!experrors)
{
System.Console.Write("OK");
System.Console.WriteLine();
}
else
{
System.Console.Write("FAILED");
System.Console.WriteLine();
}
if (waserrors)
{
System.Console.Write("TEST SUMMARY: FAILED");
System.Console.WriteLine();
}
else
{
System.Console.Write("TEST SUMMARY: PASSED");
System.Console.WriteLine();
}
System.Console.WriteLine();
System.Console.WriteLine();
}
result = !waserrors;
return(result);
}
public static bool testhqrnd(bool silent)
{
bool result = new bool();
bool waserrors = new bool();
int samplesize = 0;
double sigmathreshold = 0;
int passcount = 0;
int n = 0;
int i = 0;
int pass = 0;
int s1 = 0;
int s2 = 0;
int i1 = 0;
int i2 = 0;
double r1 = 0;
double r2 = 0;
double[] x = new double[0];
double mean = 0;
double means = 0;
double stddev = 0;
double stddevs = 0;
double lambda = 0;
bool seederrors = new bool();
bool urerrors = new bool();
double ursigmaerr = 0;
bool uierrors = new bool();
double uisigmaerr = 0;
bool normerrors = new bool();
double normsigmaerr = 0;
bool experrors = new bool();
double expsigmaerr = 0;
hqrnd.hqrndstate state = new hqrnd.hqrndstate();
waserrors = false;
sigmathreshold = 7;
samplesize = 100000;
passcount = 50;
x = new double[samplesize-1+1];
//
// Test seed errors
//
seederrors = false;
for(pass=1; pass<=passcount; pass++)
{
s1 = 1+AP.Math.RandomInteger(32000);
s2 = 1+AP.Math.RandomInteger(32000);
unsetstate(ref state);
hqrnd.hqrndseed(s1, s2, ref state);
i1 = hqrnd.hqrnduniformi(100, ref state);
unsetstate(ref state);
hqrnd.hqrndseed(s1, s2, ref state);
i2 = hqrnd.hqrnduniformi(100, ref state);
seederrors = seederrors | i1!=i2;
unsetstate(ref state);
hqrnd.hqrndseed(s1, s2, ref state);
r1 = hqrnd.hqrnduniformr(ref state);
unsetstate(ref state);
hqrnd.hqrndseed(s1, s2, ref state);
r2 = hqrnd.hqrnduniformr(ref state);
seederrors = seederrors | (double)(r1)!=(double)(r2);
}
//
// Test HQRNDRandomize() and real uniform generator
//
unsetstate(ref state);
hqrnd.hqrndrandomize(ref state);
urerrors = false;
ursigmaerr = 0;
for(i=0; i<=samplesize-1; i++)
{
x[i] = hqrnd.hqrnduniformr(ref state);
}
for(i=0; i<=samplesize-1; i++)
{
urerrors = urerrors | (double)(x[i])<=(double)(0) | (double)(x[i])>=(double)(1);
}
calculatemv(ref x, samplesize, ref mean, ref means, ref stddev, ref stddevs);
if( (double)(means)!=(double)(0) )
{
ursigmaerr = Math.Max(ursigmaerr, Math.Abs((mean-0.5)/means));
}
else
{
urerrors = true;
}
if( (double)(stddevs)!=(double)(0) )
{
ursigmaerr = Math.Max(ursigmaerr, Math.Abs((stddev-Math.Sqrt((double)(1)/(double)(12)))/stddevs));
}
else
{
urerrors = true;
}
urerrors = urerrors | (double)(ursigmaerr)>(double)(sigmathreshold);
//
// Test HQRNDRandomize() and integer uniform
//
unsetstate(ref state);
hqrnd.hqrndrandomize(ref state);
uierrors = false;
uisigmaerr = 0;
for(n=2; n<=10; n++)
{
for(i=0; i<=samplesize-1; i++)
{
x[i] = hqrnd.hqrnduniformi(n, ref state);
}
for(i=0; i<=samplesize-1; i++)
{
uierrors = uierrors | (double)(x[i])<(double)(0) | (double)(x[i])>=(double)(n);
}
calculatemv(ref x, samplesize, ref mean, ref means, ref stddev, ref stddevs);
if( (double)(means)!=(double)(0) )
{
uisigmaerr = Math.Max(uisigmaerr, Math.Abs((mean-0.5*(n-1))/means));
}
else
{
uierrors = true;
}
if( (double)(stddevs)!=(double)(0) )
{
uisigmaerr = Math.Max(uisigmaerr, Math.Abs((stddev-Math.Sqrt((AP.Math.Sqr(n)-1)/12))/stddevs));
}
else
{
uierrors = true;
}
}
uierrors = uierrors | (double)(uisigmaerr)>(double)(sigmathreshold);
//
// Special 'close-to-limit' test on uniformity of integers
// (straightforward implementation like 'RND mod N' will return
// non-uniform numbers for N=2/3*LIMIT)
//
unsetstate(ref state);
hqrnd.hqrndrandomize(ref state);
uierrors = false;
uisigmaerr = 0;
n = (int)Math.Round(2.0/3.0*2147483563.0);
for(i=0; i<=samplesize-1; i++)
{
x[i] = hqrnd.hqrnduniformi(n, ref state);
}
for(i=0; i<=samplesize-1; i++)
{
uierrors = uierrors | (double)(x[i])<(double)(0) | (double)(x[i])>=(double)(n);
}
calculatemv(ref x, samplesize, ref mean, ref means, ref stddev, ref stddevs);
if( (double)(means)!=(double)(0) )
{
uisigmaerr = Math.Max(uisigmaerr, Math.Abs((mean-0.5*(n-1))/means));
}
else
{
uierrors = true;
}
if( (double)(stddevs)!=(double)(0) )
{
uisigmaerr = Math.Max(uisigmaerr, Math.Abs((stddev-Math.Sqrt((AP.Math.Sqr(n)-1)/12))/stddevs));
}
else
{
uierrors = true;
}
uierrors = uierrors | (double)(uisigmaerr)>(double)(sigmathreshold);
//
// Test normal
//
unsetstate(ref state);
hqrnd.hqrndrandomize(ref state);
normerrors = false;
normsigmaerr = 0;
i = 0;
while( i<samplesize )
{
hqrnd.hqrndnormal2(ref state, ref r1, ref r2);
x[i] = r1;
if( i+1<samplesize )
{
x[i+1] = r2;
}
i = i+2;
}
calculatemv(ref x, samplesize, ref mean, ref means, ref stddev, ref stddevs);
if( (double)(means)!=(double)(0) )
{
normsigmaerr = Math.Max(normsigmaerr, Math.Abs((mean-0)/means));
}
else
{
normerrors = true;
}
if( (double)(stddevs)!=(double)(0) )
{
normsigmaerr = Math.Max(normsigmaerr, Math.Abs((stddev-1)/stddevs));
}
else
{
normerrors = true;
}
normerrors = normerrors | (double)(normsigmaerr)>(double)(sigmathreshold);
//
// Test exponential
//
unsetstate(ref state);
hqrnd.hqrndrandomize(ref state);
experrors = false;
expsigmaerr = 0;
lambda = 2+5*AP.Math.RandomReal();
for(i=0; i<=samplesize-1; i++)
{
x[i] = hqrnd.hqrndexponential(lambda, ref state);
}
for(i=0; i<=samplesize-1; i++)
{
uierrors = uierrors | (double)(x[i])<(double)(0);
}
calculatemv(ref x, samplesize, ref mean, ref means, ref stddev, ref stddevs);
if( (double)(means)!=(double)(0) )
{
expsigmaerr = Math.Max(expsigmaerr, Math.Abs((mean-1.0/lambda)/means));
}
else
{
experrors = true;
}
if( (double)(stddevs)!=(double)(0) )
{
expsigmaerr = Math.Max(expsigmaerr, Math.Abs((stddev-1.0/lambda)/stddevs));
}
else
{
experrors = true;
}
experrors = experrors | (double)(expsigmaerr)>(double)(sigmathreshold);
//
// Final report
//
waserrors = seederrors | urerrors | uierrors | normerrors | experrors;
if( !silent )
{
System.Console.Write("RNG TEST");
System.Console.WriteLine();
System.Console.Write("SEED TEST: ");
if( !seederrors )
{
System.Console.Write("OK");
System.Console.WriteLine();
}
else
{
System.Console.Write("FAILED");
System.Console.WriteLine();
}
System.Console.Write("UNIFORM CONTINUOUS: ");
if( !urerrors )
{
System.Console.Write("OK");
System.Console.WriteLine();
}
else
{
System.Console.Write("FAILED");
System.Console.WriteLine();
}
System.Console.Write("UNIFORM INTEGER: ");
if( !uierrors )
{
System.Console.Write("OK");
System.Console.WriteLine();
}
else
{
System.Console.Write("FAILED");
System.Console.WriteLine();
}
System.Console.Write("NORMAL: ");
if( !normerrors )
{
System.Console.Write("OK");
System.Console.WriteLine();
}
else
{
System.Console.Write("FAILED");
System.Console.WriteLine();
}
System.Console.Write("EXPONENTIAL: ");
if( !experrors )
{
System.Console.Write("OK");
System.Console.WriteLine();
}
else
{
System.Console.Write("FAILED");
System.Console.WriteLine();
}
if( waserrors )
{
System.Console.Write("TEST SUMMARY: FAILED");
System.Console.WriteLine();
}
else
{
System.Console.Write("TEST SUMMARY: PASSED");
System.Console.WriteLine();
}
System.Console.WriteLine();
System.Console.WriteLine();
}
result = !waserrors;
return result;
}
/*************************************************************************
Hermitian multiplication of NxN matrix by random Haar distributed
complex orthogonal matrix
INPUT PARAMETERS:
A - matrix, array[0..N-1, 0..N-1]
N - matrix size
OUTPUT PARAMETERS:
A - Q^H*A*Q, where Q is random NxN orthogonal matrix
-- ALGLIB routine --
04.12.2009
Bochkanov Sergey
*************************************************************************/
public static void hmatrixrndmultiply(ref AP.Complex[,] a,
int n)
{
AP.Complex tau = 0;
AP.Complex lambda = 0;
int s = 0;
int i = 0;
AP.Complex[] w = new AP.Complex[0];
AP.Complex[] v = new AP.Complex[0];
hqrnd.hqrndstate state = new hqrnd.hqrndstate();
int i_ = 0;
//
// General case.
//
w = new AP.Complex[n-1+1];
v = new AP.Complex[n+1];
hqrnd.hqrndrandomize(ref state);
for(s=2; s<=n; s++)
{
//
// Prepare random normal v
//
do
{
for(i=1; i<=s; i++)
{
hqrnd.hqrndnormal2(ref state, ref tau.x, ref tau.y);
v[i] = tau;
}
lambda = 0.0;
for(i_=1; i_<=s;i_++)
{
lambda += v[i_]*AP.Math.Conj(v[i_]);
}
}
while( lambda==0 );
//
// Prepare and apply reflection
//
creflections.complexgeneratereflection(ref v, s, ref tau);
v[1] = 1;
creflections.complexapplyreflectionfromtheright(ref a, tau, ref v, 0, n-1, n-s, n-1, ref w);
creflections.complexapplyreflectionfromtheleft(ref a, AP.Math.Conj(tau), ref v, n-s, n-1, 0, n-1, ref w);
}
//
// Second pass.
//
for(i=0; i<=n-1; i++)
{
hqrnd.hqrndunit2(ref state, ref tau.x, ref tau.y);
for(i_=0; i_<=n-1;i_++)
{
a[i_,i] = tau*a[i_,i];
}
tau = AP.Math.Conj(tau);
for(i_=0; i_<=n-1;i_++)
{
a[i,i_] = tau*a[i,i_];
}
}
}
/*************************************************************************
Symmetric multiplication of NxN matrix by random Haar distributed
orthogonal matrix
INPUT PARAMETERS:
A - matrix, array[0..N-1, 0..N-1]
N - matrix size
OUTPUT PARAMETERS:
A - Q'*A*Q, where Q is random NxN orthogonal matrix
-- ALGLIB routine --
04.12.2009
Bochkanov Sergey
*************************************************************************/
public static void smatrixrndmultiply(ref double[,] a,
int n)
{
double tau = 0;
double lambda = 0;
int s = 0;
int i = 0;
double u1 = 0;
double u2 = 0;
double[] w = new double[0];
double[] v = new double[0];
hqrnd.hqrndstate state = new hqrnd.hqrndstate();
int i_ = 0;
//
// General case.
//
w = new double[n-1+1];
v = new double[n+1];
hqrnd.hqrndrandomize(ref state);
for(s=2; s<=n; s++)
{
//
// Prepare random normal v
//
do
{
i = 1;
while( i<=s )
{
hqrnd.hqrndnormal2(ref state, ref u1, ref u2);
v[i] = u1;
if( i+1<=s )
{
v[i+1] = u2;
}
i = i+2;
}
lambda = 0.0;
for(i_=1; i_<=s;i_++)
{
lambda += v[i_]*v[i_];
}
}
while( (double)(lambda)==(double)(0) );
//
// Prepare and apply reflection
//
reflections.generatereflection(ref v, s, ref tau);
v[1] = 1;
reflections.applyreflectionfromtheright(ref a, tau, ref v, 0, n-1, n-s, n-1, ref w);
reflections.applyreflectionfromtheleft(ref a, tau, ref v, n-s, n-1, 0, n-1, ref w);
}
//
// Second pass.
//
for(i=0; i<=n-1; i++)
{
tau = 2*AP.Math.RandomInteger(2)-1;
for(i_=0; i_<=n-1;i_++)
{
a[i_,i] = tau*a[i_,i];
}
for(i_=0; i_<=n-1;i_++)
{
a[i,i_] = tau*a[i,i_];
}
}
}
/*************************************************************************
Multiplication of MxN complex matrix by MxM random Haar distributed
complex orthogonal matrix
INPUT PARAMETERS:
A - matrix, array[0..M-1, 0..N-1]
M, N- matrix size
OUTPUT PARAMETERS:
A - Q*A, where Q is random MxM orthogonal matrix
-- ALGLIB routine --
04.12.2009
Bochkanov Sergey
*************************************************************************/
public static void cmatrixrndorthogonalfromtheleft(ref AP.Complex[,] a,
int m,
int n)
{
AP.Complex tau = 0;
AP.Complex lambda = 0;
int s = 0;
int i = 0;
int j = 0;
AP.Complex[] w = new AP.Complex[0];
AP.Complex[] v = new AP.Complex[0];
hqrnd.hqrndstate state = new hqrnd.hqrndstate();
int i_ = 0;
System.Diagnostics.Debug.Assert(n>=1 & m>=1, "CMatrixRndOrthogonalFromTheRight: N<1 or M<1!");
if( m==1 )
{
//
// special case
//
hqrnd.hqrndrandomize(ref state);
hqrnd.hqrndunit2(ref state, ref tau.x, ref tau.y);
for(j=0; j<=n-1; j++)
{
a[0,j] = a[0,j]*tau;
}
return;
}
//
// General case.
// First pass.
//
w = new AP.Complex[n-1+1];
v = new AP.Complex[m+1];
hqrnd.hqrndrandomize(ref state);
for(s=2; s<=m; s++)
{
//
// Prepare random normal v
//
do
{
for(i=1; i<=s; i++)
{
hqrnd.hqrndnormal2(ref state, ref tau.x, ref tau.y);
v[i] = tau;
}
lambda = 0.0;
for(i_=1; i_<=s;i_++)
{
lambda += v[i_]*AP.Math.Conj(v[i_]);
}
}
while( lambda==0 );
//
// Prepare and apply reflection
//
creflections.complexgeneratereflection(ref v, s, ref tau);
v[1] = 1;
creflections.complexapplyreflectionfromtheleft(ref a, tau, ref v, m-s, m-1, 0, n-1, ref w);
}
//
// Second pass.
//
for(i=0; i<=m-1; i++)
{
hqrnd.hqrndunit2(ref state, ref tau.x, ref tau.y);
for(i_=0; i_<=n-1;i_++)
{
a[i,i_] = tau*a[i,i_];
}
}
}
/*************************************************************************
Multiplication of MxN matrix by MxM random Haar distributed orthogonal matrix
INPUT PARAMETERS:
A - matrix, array[0..M-1, 0..N-1]
M, N- matrix size
OUTPUT PARAMETERS:
A - Q*A, where Q is random MxM orthogonal matrix
-- ALGLIB routine --
04.12.2009
Bochkanov Sergey
*************************************************************************/
public static void rmatrixrndorthogonalfromtheleft(ref double[,] a,
int m,
int n)
{
double tau = 0;
double lambda = 0;
int s = 0;
int i = 0;
int j = 0;
double u1 = 0;
double u2 = 0;
double[] w = new double[0];
double[] v = new double[0];
hqrnd.hqrndstate state = new hqrnd.hqrndstate();
int i_ = 0;
System.Diagnostics.Debug.Assert(n>=1 & m>=1, "RMatrixRndOrthogonalFromTheRight: N<1 or M<1!");
if( m==1 )
{
//
// special case
//
tau = 2*AP.Math.RandomInteger(2)-1;
for(j=0; j<=n-1; j++)
{
a[0,j] = a[0,j]*tau;
}
return;
}
//
// General case.
// First pass.
//
w = new double[n-1+1];
v = new double[m+1];
hqrnd.hqrndrandomize(ref state);
for(s=2; s<=m; s++)
{
//
// Prepare random normal v
//
do
{
i = 1;
while( i<=s )
{
hqrnd.hqrndnormal2(ref state, ref u1, ref u2);
v[i] = u1;
if( i+1<=s )
{
v[i+1] = u2;
}
i = i+2;
}
lambda = 0.0;
for(i_=1; i_<=s;i_++)
{
lambda += v[i_]*v[i_];
}
}
while( (double)(lambda)==(double)(0) );
//
// Prepare and apply reflection
//
reflections.generatereflection(ref v, s, ref tau);
v[1] = 1;
reflections.applyreflectionfromtheleft(ref a, tau, ref v, m-s, m-1, 0, n-1, ref w);
}
//
// Second pass.
//
for(i=0; i<=m-1; i++)
{
tau = 2*AP.Math.RandomInteger(2)-1;
for(i_=0; i_<=n-1;i_++)
{
a[i,i_] = tau*a[i,i_];
}
}
}
/*************************************************************************
Generation of random NxN complex matrix with given condition number C and
norm2(A)=1
INPUT PARAMETERS:
N - matrix size
C - condition number (in 2-norm)
OUTPUT PARAMETERS:
A - random matrix with norm2(A)=1 and cond(A)=C
-- ALGLIB routine --
04.12.2009
Bochkanov Sergey
*************************************************************************/
public static void cmatrixrndcond(int n,
double c,
ref AP.Complex[,] a)
{
int i = 0;
int j = 0;
double l1 = 0;
double l2 = 0;
hqrnd.hqrndstate state = new hqrnd.hqrndstate();
AP.Complex v = 0;
System.Diagnostics.Debug.Assert(n>=1 & (double)(c)>=(double)(1), "CMatrixRndCond: N<1 or C<1!");
a = new AP.Complex[n-1+1, n-1+1];
if( n==1 )
{
//
// special case
//
hqrnd.hqrndrandomize(ref state);
hqrnd.hqrndunit2(ref state, ref v.x, ref v.y);
a[0,0] = v;
return;
}
l1 = 0;
l2 = Math.Log(1/c);
for(i=0; i<=n-1; i++)
{
for(j=0; j<=n-1; j++)
{
a[i,j] = 0;
}
}
a[0,0] = Math.Exp(l1);
for(i=1; i<=n-2; i++)
{
a[i,i] = Math.Exp(AP.Math.RandomReal()*(l2-l1)+l1);
}
a[n-1,n-1] = Math.Exp(l2);
cmatrixrndorthogonalfromtheleft(ref a, n, n);
cmatrixrndorthogonalfromtheright(ref a, n, n);
}
/*************************************************************************
* Hermitian multiplication of NxN matrix by random Haar distributed
* complex orthogonal matrix
*
* INPUT PARAMETERS:
* A - matrix, array[0..N-1, 0..N-1]
* N - matrix size
*
* OUTPUT PARAMETERS:
* A - Q^H*A*Q, where Q is random NxN orthogonal matrix
*
* -- ALGLIB routine --
* 04.12.2009
* Bochkanov Sergey
*************************************************************************/
public static void hmatrixrndmultiply(ref AP.Complex[,] a,
int n)
{
AP.Complex tau = 0;
AP.Complex lambda = 0;
int s = 0;
int i = 0;
AP.Complex[] w = new AP.Complex[0];
AP.Complex[] v = new AP.Complex[0];
hqrnd.hqrndstate state = new hqrnd.hqrndstate();
int i_ = 0;
//
// General case.
//
w = new AP.Complex[n - 1 + 1];
v = new AP.Complex[n + 1];
hqrnd.hqrndrandomize(ref state);
for (s = 2; s <= n; s++)
{
//
// Prepare random normal v
//
do
{
for (i = 1; i <= s; i++)
{
hqrnd.hqrndnormal2(ref state, ref tau.x, ref tau.y);
v[i] = tau;
}
lambda = 0.0;
for (i_ = 1; i_ <= s; i_++)
{
lambda += v[i_] * AP.Math.Conj(v[i_]);
}
}while(lambda == 0);
//
// Prepare and apply reflection
//
creflections.complexgeneratereflection(ref v, s, ref tau);
v[1] = 1;
creflections.complexapplyreflectionfromtheright(ref a, tau, ref v, 0, n - 1, n - s, n - 1, ref w);
creflections.complexapplyreflectionfromtheleft(ref a, AP.Math.Conj(tau), ref v, n - s, n - 1, 0, n - 1, ref w);
}
//
// Second pass.
//
for (i = 0; i <= n - 1; i++)
{
hqrnd.hqrndunit2(ref state, ref tau.x, ref tau.y);
for (i_ = 0; i_ <= n - 1; i_++)
{
a[i_, i] = tau * a[i_, i];
}
tau = AP.Math.Conj(tau);
for (i_ = 0; i_ <= n - 1; i_++)
{
a[i, i_] = tau * a[i, i_];
}
}
}
/*************************************************************************
* Symmetric multiplication of NxN matrix by random Haar distributed
* orthogonal matrix
*
* INPUT PARAMETERS:
* A - matrix, array[0..N-1, 0..N-1]
* N - matrix size
*
* OUTPUT PARAMETERS:
* A - Q'*A*Q, where Q is random NxN orthogonal matrix
*
* -- ALGLIB routine --
* 04.12.2009
* Bochkanov Sergey
*************************************************************************/
public static void smatrixrndmultiply(ref double[,] a,
int n)
{
double tau = 0;
double lambda = 0;
int s = 0;
int i = 0;
double u1 = 0;
double u2 = 0;
double[] w = new double[0];
double[] v = new double[0];
hqrnd.hqrndstate state = new hqrnd.hqrndstate();
int i_ = 0;
//
// General case.
//
w = new double[n - 1 + 1];
v = new double[n + 1];
hqrnd.hqrndrandomize(ref state);
for (s = 2; s <= n; s++)
{
//
// Prepare random normal v
//
do
{
i = 1;
while (i <= s)
{
hqrnd.hqrndnormal2(ref state, ref u1, ref u2);
v[i] = u1;
if (i + 1 <= s)
{
v[i + 1] = u2;
}
i = i + 2;
}
lambda = 0.0;
for (i_ = 1; i_ <= s; i_++)
{
lambda += v[i_] * v[i_];
}
}while((double)(lambda) == (double)(0));
//
// Prepare and apply reflection
//
reflections.generatereflection(ref v, s, ref tau);
v[1] = 1;
reflections.applyreflectionfromtheright(ref a, tau, ref v, 0, n - 1, n - s, n - 1, ref w);
reflections.applyreflectionfromtheleft(ref a, tau, ref v, n - s, n - 1, 0, n - 1, ref w);
}
//
// Second pass.
//
for (i = 0; i <= n - 1; i++)
{
tau = 2 * AP.Math.RandomInteger(2) - 1;
for (i_ = 0; i_ <= n - 1; i_++)
{
a[i_, i] = tau * a[i_, i];
}
for (i_ = 0; i_ <= n - 1; i_++)
{
a[i, i_] = tau * a[i, i_];
}
}
}
/*************************************************************************
* Multiplication of MxN complex matrix by MxM random Haar distributed
* complex orthogonal matrix
*
* INPUT PARAMETERS:
* A - matrix, array[0..M-1, 0..N-1]
* M, N- matrix size
*
* OUTPUT PARAMETERS:
* A - Q*A, where Q is random MxM orthogonal matrix
*
* -- ALGLIB routine --
* 04.12.2009
* Bochkanov Sergey
*************************************************************************/
public static void cmatrixrndorthogonalfromtheleft(ref AP.Complex[,] a,
int m,
int n)
{
AP.Complex tau = 0;
AP.Complex lambda = 0;
int s = 0;
int i = 0;
int j = 0;
AP.Complex[] w = new AP.Complex[0];
AP.Complex[] v = new AP.Complex[0];
hqrnd.hqrndstate state = new hqrnd.hqrndstate();
int i_ = 0;
System.Diagnostics.Debug.Assert(n >= 1 & m >= 1, "CMatrixRndOrthogonalFromTheRight: N<1 or M<1!");
if (m == 1)
{
//
// special case
//
hqrnd.hqrndrandomize(ref state);
hqrnd.hqrndunit2(ref state, ref tau.x, ref tau.y);
for (j = 0; j <= n - 1; j++)
{
a[0, j] = a[0, j] * tau;
}
return;
}
//
// General case.
// First pass.
//
w = new AP.Complex[n - 1 + 1];
v = new AP.Complex[m + 1];
hqrnd.hqrndrandomize(ref state);
for (s = 2; s <= m; s++)
{
//
// Prepare random normal v
//
do
{
for (i = 1; i <= s; i++)
{
hqrnd.hqrndnormal2(ref state, ref tau.x, ref tau.y);
v[i] = tau;
}
lambda = 0.0;
for (i_ = 1; i_ <= s; i_++)
{
lambda += v[i_] * AP.Math.Conj(v[i_]);
}
}while(lambda == 0);
//
// Prepare and apply reflection
//
creflections.complexgeneratereflection(ref v, s, ref tau);
v[1] = 1;
creflections.complexapplyreflectionfromtheleft(ref a, tau, ref v, m - s, m - 1, 0, n - 1, ref w);
}
//
// Second pass.
//
for (i = 0; i <= m - 1; i++)
{
hqrnd.hqrndunit2(ref state, ref tau.x, ref tau.y);
for (i_ = 0; i_ <= n - 1; i_++)
{
a[i, i_] = tau * a[i, i_];
}
}
}
/*************************************************************************
* Multiplication of MxN matrix by MxM random Haar distributed orthogonal matrix
*
* INPUT PARAMETERS:
* A - matrix, array[0..M-1, 0..N-1]
* M, N- matrix size
*
* OUTPUT PARAMETERS:
* A - Q*A, where Q is random MxM orthogonal matrix
*
* -- ALGLIB routine --
* 04.12.2009
* Bochkanov Sergey
*************************************************************************/
public static void rmatrixrndorthogonalfromtheleft(ref double[,] a,
int m,
int n)
{
double tau = 0;
double lambda = 0;
int s = 0;
int i = 0;
int j = 0;
double u1 = 0;
double u2 = 0;
double[] w = new double[0];
double[] v = new double[0];
hqrnd.hqrndstate state = new hqrnd.hqrndstate();
int i_ = 0;
System.Diagnostics.Debug.Assert(n >= 1 & m >= 1, "RMatrixRndOrthogonalFromTheRight: N<1 or M<1!");
if (m == 1)
{
//
// special case
//
tau = 2 * AP.Math.RandomInteger(2) - 1;
for (j = 0; j <= n - 1; j++)
{
a[0, j] = a[0, j] * tau;
}
return;
}
//
// General case.
// First pass.
//
w = new double[n - 1 + 1];
v = new double[m + 1];
hqrnd.hqrndrandomize(ref state);
for (s = 2; s <= m; s++)
{
//
// Prepare random normal v
//
do
{
i = 1;
while (i <= s)
{
hqrnd.hqrndnormal2(ref state, ref u1, ref u2);
v[i] = u1;
if (i + 1 <= s)
{
v[i + 1] = u2;
}
i = i + 2;
}
lambda = 0.0;
for (i_ = 1; i_ <= s; i_++)
{
lambda += v[i_] * v[i_];
}
}while((double)(lambda) == (double)(0));
//
// Prepare and apply reflection
//
reflections.generatereflection(ref v, s, ref tau);
v[1] = 1;
reflections.applyreflectionfromtheleft(ref a, tau, ref v, m - s, m - 1, 0, n - 1, ref w);
}
//
// Second pass.
//
for (i = 0; i <= m - 1; i++)
{
tau = 2 * AP.Math.RandomInteger(2) - 1;
for (i_ = 0; i_ <= n - 1; i_++)
{
a[i, i_] = tau * a[i, i_];
}
}
}
