README

@---------------------------------------------------------------------------
@---------------------------------------------------------------------------
@ FILE: dsylv.c
@---------------------------------------------------------------------------
@---------------------------------------------------------------------------
@
@---------------------------------------------------------------------------
@ matrix_TYP *dsylv(M)
@ matrix_TYP *M;
@ 
@ The return of 'dsylv' is a diagonal matrix D with diaogonal elements
@ D[i][i] in {-1,1,0}.
@ The function diagonalizes the symmetric matrix 'M' by simultanous row
@ and col-operations.
@ If the entry in the diagonalised matrix is positiv it is replaced
@ by 1, if negativ by -1, if zero by 0.
@ This is done with floating-point arithmetic.
@ So rounding errows may occur.
@ Therefore a diaogonal entry it is assumed to be zero
@ if its absolute value is less then eps = 0.000001
@
@---------------------------------------------------------------------------
@
@---------------------------------------------------------------------------
@ int definite_test(M)
@ matrix_TYP *M;
@ 
@ Does the same as 'dsolve' but testes during the Diagonalisation
@ if the matrix is indefinite.
@ The return is
@ 
@      0:  if M is zero
@      1:  if M is positiv semidefinite, but not positiv definite.
@      2:  if M is positiv definite.
@     -1:  if M is negativ semidefinite, but not negativ definite.
@     -2:  if M is negativ definite.
@     -3:  if M is indefinite.
@
@---------------------------------------------------------------------------
@
@---------------------------------------------------------------------------
@---------------------------------------------------------------------------
@ FILE: rest_short.c
@---------------------------------------------------------------------------
@---------------------------------------------------------------------------
@
@---------------------------------------------------------------------------
@ matrix_TYP *rest_short(mat, restvec, rkgv, zaehler, nenner,
@                        find_opt, count_opt, anz)
@ matrix_TYP *mat;
@ int *restvec, rkgv;
@ int zaehler, nenner, find_opt, count_opt, *anz;
@
@ The matrix mat must be a symmetric, positiv definite nxn-Matrix.
@ The function calculates all x in Z^n  with 
@ 
@          lengthmin <= n(x - 1/rkgv * restvec)  <= length
@ where
@          n(y) := y * mat * y^{tr}.
@ 
@ In the pointer 'anz' the number of all these vectors is returned.
@ The vectors x are the rows of the returned matrix, called 'SV' in the
@ following.
@ So 'SV' is an integral matrix with 'anz' rows and 'n' columns.
@ For 'SV' n+1 columns are allocated. The last column contains the norm n(x).
@ Although n+1 columns are allocated, SV->cols = n.
@ 
@ If 'findoption == 1' the Algorithm stops,
@ if a vector of desired norm is found.
@ In this case the Matrix 'SV' has only one row.
@ 
@ If 'count_option == 1',  shortest_vectors returns a zero-pointer,
@ only the number of found vectors is returned in the pointer 'anz'.
@ 
@ Warning: if 'mat' is not positiv definite, the function runs into an 
@          infinite loop.
@
@---------------------------------------------------------------------------
@
@---------------------------------------------------------------------------
@---------------------------------------------------------------------------
@ FILE: short.c
@---------------------------------------------------------------------------
@---------------------------------------------------------------------------
@
@---------------------------------------------------------------------------
@ matrix_TYP *short_vectors(mat, length, lengthmin, find_opt, count_opt, anz)
@ matrix_TYP *mat;
@ int length, lengthmin, find_opt, count_opt, *anz;
@ 
@ The matrix mat must be a symmetric, positiv definite nxn-Matrix.
@ The function calculates all x in Z^n ( up to multiplication with -1) with 
@ 
@          lengthmin <= n(x)  <= length
@ where
@          n(x) := x * mat * x^{tr}.
@ 
@ In the pointer 'anz' the number of all these vectors is returned.
@ The vectors x are the rows of the returned matrix,
@ called 'SV' in the following.
@ So 'SV' is an integral matrix with 'anz' rows and 'n' columns.
@ For 'SV' n+1 columns are allocated. The last column contains the norm n(x).
@ Although n+1 columns are allocated, SV->cols = n.
@ 
@ If 'findoption == 1' the Algorithm stops,
@ if a vector of desired norm is found.
@ In this case the Matrix 'SV' has only one row.
@ 
@ If 'count_option == 1',  shortest_vectors returns a zero-pointer,
@ only the number of found vectors is returned in the pointer 'anz'.
@ 
@ Warning: if 'mat' is not positiv definite, the function runs into an 
@          infinite loop.
@
@---------------------------------------------------------------------------
@
@---------------------------------------------------------------------------
@---------------------------------------------------------------------------
@ FILE: shortest.c
@---------------------------------------------------------------------------
@---------------------------------------------------------------------------
@
@---------------------------------------------------------------------------
@ matrix_TYP *shortest(mat, min_norm)
@ matrix_TYP *mat;
@ int  *min_norm;
@ 
@ The function callulates the shortest vectors of the integral, symmetric,
@ positiv definite matrix 'mat'.
@ The minmum
@ 
@    m(mat) := min{  v * mat *v^{tr} | v in Z^n and v not 0}
@ 
@ is returned in the pointer 'min_norm'. The shortest vectors are the vectors
@ with n(x) = min(x).
@ They are stored as rows in the returned matrix called 'SV' in the following.
@ For 'SV' n+1 columns are allocated. The last column contains the norm n(x).
@ Although n+1 columns are allocated, SV->cols = n.
@ 
@ Warning: if 'mat' is not positiv definite, the function runs into an 
@         infinite loop.
@---------------------------------------------------------------------------
@