func updateScaling(W *FloatMatrixSet, lmbda, s, z *matrix.FloatMatrix) (err error) {
err = nil
var stmp, ztmp *matrix.FloatMatrix
/*
Nonlinear and 'l' blocks
d := d .* sqrt( s ./ z )
lmbda := lmbda .* sqrt(s) .* sqrt(z)
*/
mnl := 0
dnlset := W.At("dnl")
dnliset := W.At("dnli")
dset := W.At("d")
diset := W.At("di")
beta := W.At("beta")[0]
if dnlset != nil && dnlset[0].NumElements() > 0 {
mnl = dnlset[0].NumElements()
}
ml := dset[0].NumElements()
m := mnl + ml
//fmt.Printf("ml=%d, mnl=%d, m=%d'n", ml, mnl, m)
stmp = matrix.FloatVector(s.FloatArray()[:m])
stmp.Apply(stmp, math.Sqrt)
s.SetIndexes(matrix.MakeIndexSet(0, m, 1), stmp.FloatArray())
ztmp = matrix.FloatVector(z.FloatArray()[:m])
ztmp.Apply(ztmp, math.Sqrt)
z.SetIndexes(matrix.MakeIndexSet(0, m, 1), ztmp.FloatArray())
// d := d .* s .* z
if len(dnlset) > 0 {
blas.TbmvFloat(s, dnlset[0], &la_.IOpt{"n", mnl}, &la_.IOpt{"k", 0}, &la_.IOpt{"lda", 1})
blas.TbsvFloat(z, dnlset[0], &la_.IOpt{"n", mnl}, &la_.IOpt{"k", 0}, &la_.IOpt{"lda", 1})
dnliset[0].Apply(dnlset[0], func(a float64) float64 { return 1.0 / a })
}
blas.TbmvFloat(s, dset[0], &la_.IOpt{"n", ml},
&la_.IOpt{"k", 0}, &la_.IOpt{"lda", 1}, &la_.IOpt{"offseta", mnl})
blas.TbsvFloat(z, dset[0], &la_.IOpt{"n", ml},
&la_.IOpt{"k", 0}, &la_.IOpt{"lda", 1}, &la_.IOpt{"offseta", mnl})
diset[0].Apply(dset[0], func(a float64) float64 { return 1.0 / a })
// lmbda := s .* z
blas.CopyFloat(s, lmbda, &la_.IOpt{"n", m})
blas.TbmvFloat(z, lmbda, &la_.IOpt{"n", m}, &la_.IOpt{"k", 0}, &la_.IOpt{"lda", 1})
//fmt.Printf("-- end of l:\nz=\n%v\nlmbda=\n%v\n", z.ConvertToString(), lmbda.ConvertToString())
//fmt.Printf("W[d]=\n%v\n", dset[0].ConvertToString())
//fmt.Printf("W[di]=\n%v\n", diset[0].ConvertToString())
// 'q' blocks.
// Let st and zt be the new variables in the old scaling:
//
// st = s_k, zt = z_k
//
// and a = sqrt(st' * J * st), b = sqrt(zt' * J * zt).
//
// 1. Compute the hyperbolic Householder transformation 2*q*q' - J
// that maps st/a to zt/b.
//
// c = sqrt( (1 + st'*zt/(a*b)) / 2 )
// q = (st/a + J*zt/b) / (2*c).
//
// The new scaling point is
//
// wk := betak * sqrt(a/b) * (2*v[k]*v[k]' - J) * q
//
// with betak = W['beta'][k].
//
// 3. The scaled variable:
//
// lambda_k0 = sqrt(a*b) * c
// lambda_k1 = sqrt(a*b) * ( (2vk*vk' - J) * (-d*q + u/2) )_1
//
// where
//
// u = st/a - J*zt/b
// d = ( vk0 * (vk'*u) + u0/2 ) / (2*vk0 *(vk'*q) - q0 + 1).
//
// 4. Update scaling
//
// v[k] := wk^1/2
// = 1 / sqrt(2*(wk0 + 1)) * (wk + e).
// beta[k] *= sqrt(a/b)
ind := m
for k, v := range W.At("v") {
m = v.NumElements()
// ln = sqrt( lambda_k' * J * lambda_k ) !! NOT USED!!
jnrm2(lmbda, m, ind) // ?? NOT USED ??
// a = sqrt( sk' * J * sk ) = sqrt( st' * J * st )
// s := s / a = st / a
aa := jnrm2(s, m, ind)
blas.ScalFloat(s, 1.0/aa, &la_.IOpt{"n", m}, &la_.IOpt{"offset", ind})
// b = sqrt( zk' * J * zk ) = sqrt( zt' * J * zt )
// z := z / a = zt / b
bb := jnrm2(z, m, ind)
blas.ScalFloat(z, 1.0/bb, &la_.IOpt{"n", m}, &la_.IOpt{"offset", ind})
// c = sqrt( ( 1 + (st'*zt) / (a*b) ) / 2 )
cc := blas.DotFloat(s, z, &la_.IOpt{"offsetx", ind}, &la_.IOpt{"offsety", ind},
&la_.IOpt{"n", m})
cc = math.Sqrt((1.0 + cc) / 2.0)
// vs = v' * st / a
vs := blas.DotFloat(v, s, &la_.IOpt{"offsety", ind}, &la_.IOpt{"n", m})
// vz = v' * J *zt / b
vz := jdot(v, z, m, 0, ind)
// vq = v' * q where q = (st/a + J * zt/b) / (2 * c)
vq := (vs + vz) / 2.0 / cc
// vq = v' * q where q = (st/a + J * zt/b) / (2 * c)
vu := vs - vz
// lambda_k0 = c
lmbda.SetIndex(ind, cc)
// wk0 = 2 * vk0 * (vk' * q) - q0
wk0 := 2.0*v.GetIndex(0)*vq - (s.GetIndex(ind)+z.GetIndex(ind))/2.0/cc
// d = (v[0] * (vk' * u) - u0/2) / (wk0 + 1)
dd := (v.GetIndex(0)*vu - s.GetIndex(ind)/2.0 + z.GetIndex(ind)/2.0) / (wk0 + 1.0)
// lambda_k1 = 2 * v_k1 * vk' * (-d*q + u/2) - d*q1 + u1/2
blas.CopyFloat(v, lmbda, &la_.IOpt{"offsetx", 1}, &la_.IOpt{"offsety", ind + 1},
&la_.IOpt{"n", m - 1})
blas.ScalFloat(lmbda, (2.0 * (-dd*vq + 0.5*vu)),
&la_.IOpt{"offsetx", ind + 1}, &la_.IOpt{"offsety", ind + 1}, &la_.IOpt{"n", m - 1})
blas.AxpyFloat(s, lmbda, 0.5*(1.0-dd/cc),
&la_.IOpt{"offsetx", ind + 1}, &la_.IOpt{"offsety", ind + 1}, &la_.IOpt{"n", m - 1})
blas.AxpyFloat(z, lmbda, 0.5*(1.0+dd/cc),
&la_.IOpt{"offsetx", ind + 1}, &la_.IOpt{"offsety", ind + 1}, &la_.IOpt{"n", m - 1})
// Scale so that sqrt(lambda_k' * J * lambda_k) = sqrt(aa*bb).
blas.ScalFloat(lmbda, math.Sqrt(aa*bb), &la_.IOpt{"offset", ind}, &la_.IOpt{"n", m})
// v := (2*v*v' - J) * q
// = 2 * (v'*q) * v' - (J* st/a + zt/b) / (2*c)
blas.ScalFloat(v, 2.0*vq)
v.SetIndex(0, v.GetIndex(0)-(s.GetIndex(ind)/2.0/cc))
blas.AxpyFloat(s, v, 0.5/cc, &la_.IOpt{"offsetx", ind + 1}, &la_.IOpt{"offsety", 1},
&la_.IOpt{"n", m - 1})
blas.AxpyFloat(z, v, -0.5/cc, &la_.IOpt{"offsetx", ind}, &la_.IOpt{"n", m})
// v := v^{1/2} = 1/sqrt(2 * (v0 + 1)) * (v + e)
v0 := v.GetIndex(0) + 1.0
v.SetIndex(0, v0)
blas.ScalFloat(v, 1.0/math.Sqrt(2.0*v0))
// beta[k] *= ( aa / bb )**1/2
bk := beta.GetIndex(k)
beta.SetIndex(k, bk*math.Sqrt(aa/bb))
ind += m
}
//fmt.Printf("-- end of q:\nz=\n%v\nlmbda=\n%v\n", z.ConvertToString(), lmbda.ConvertToString())
//fmt.Printf("beta=\n%v\n", beta.ConvertToString())
// 's' blocks
//
// Let st, zt be the updated variables in the old scaling:
//
// st = Ls * Ls', zt = Lz * Lz'.
//
// where Ls and Lz are the 's' components of s, z.
//
// 1. SVD Lz'*Ls = Uk * lambda_k^+ * Vk'.
//
// 2. New scaling is
//
// r[k] := r[k] * Ls * Vk * diag(lambda_k^+)^{-1/2}
// rti[k] := r[k] * Lz * Uk * diag(lambda_k^+)^{-1/2}.
//
maxr := 0
for _, m := range W.At("r") {
if m.Rows() > maxr {
maxr = m.Rows()
}
}
work := matrix.FloatZeros(maxr*maxr, 1)
vlensum := 0
for _, m := range W.At("v") {
vlensum += m.NumElements()
}
ind = mnl + ml + vlensum
ind2 := ind
ind3 := 0
rset := W.At("r")
rtiset := W.At("rti")
for k, _ := range rset {
r := rset[k]
rti := rtiset[k]
m = r.Rows()
//fmt.Printf("m=%d, r=\n%v\nrti=\n%v\n", m, r.ConvertToString(), rti.ConvertToString())
// r := r*sk = r*Ls
blas.GemmFloat(r, s, work, 1.0, 0.0, &la_.IOpt{"m", m}, &la_.IOpt{"n", m},
&la_.IOpt{"k", m}, &la_.IOpt{"ldb", m}, &la_.IOpt{"ldc", m},
&la_.IOpt{"offsetb", ind2})
//fmt.Printf("1 work=\n%v\n", work.ConvertToString())
blas.CopyFloat(work, r, &la_.IOpt{"n", m * m})
// rti := rti*zk = rti*Lz
blas.GemmFloat(rti, z, work, 1.0, 0.0, &la_.IOpt{"m", m}, &la_.IOpt{"n", m},
&la_.IOpt{"k", m}, &la_.IOpt{"ldb", m}, &la_.IOpt{"ldc", m},
&la_.IOpt{"offsetb", ind2})
//fmt.Printf("2 work=\n%v\n", work.ConvertToString())
blas.CopyFloat(work, rti, &la_.IOpt{"n", m * m})
// SVD Lz'*Ls = U * lmbds^+ * V'; store U in sk and V' in zk. '
blas.GemmFloat(z, s, work, 1.0, 0.0, la_.OptTransA, &la_.IOpt{"m", m},
&la_.IOpt{"n", m}, &la_.IOpt{"k", m}, &la_.IOpt{"lda", m}, &la_.IOpt{"ldb", m},
&la_.IOpt{"ldc", m}, &la_.IOpt{"offseta", ind2}, &la_.IOpt{"offsetb", ind2})
//fmt.Printf("3 work=\n%v\n", work.ConvertToString())
// U = s, Vt = z
lapack.GesvdFloat(work, lmbda, s, z, la_.OptJobuAll, la_.OptJobvtAll,
&la_.IOpt{"m", m}, &la_.IOpt{"n", m}, &la_.IOpt{"lda", m}, &la_.IOpt{"ldu", m},
&la_.IOpt{"ldvt", m}, &la_.IOpt{"offsets", ind}, &la_.IOpt{"offsetu", ind2},
&la_.IOpt{"offsetvt", ind2})
// r := r*V
blas.GemmFloat(r, z, work, 1.0, 0.0, la_.OptTransB, &la_.IOpt{"m", m},
&la_.IOpt{"n", m}, &la_.IOpt{"k", m}, &la_.IOpt{"ldb", m}, &la_.IOpt{"ldc", m},
&la_.IOpt{"offsetb", ind2})
//fmt.Printf("4 work=\n%v\n", work.ConvertToString())
blas.CopyFloat(work, r, &la_.IOpt{"n", m * m})
// rti := rti*U
blas.GemmFloat(rti, s, work, 1.0, 0.0, &la_.IOpt{"m", m}, &la_.IOpt{"n", m},
&la_.IOpt{"k", m}, &la_.IOpt{"ldb", m}, &la_.IOpt{"ldc", m},
&la_.IOpt{"offsetb", ind2})
//fmt.Printf("5 work=\n%v\n", work.ConvertToString())
blas.CopyFloat(work, rti, &la_.IOpt{"n", m * m})
for i := 0; i < m; i++ {
a := 1.0 / math.Sqrt(lmbda.GetIndex(ind+i))
blas.ScalFloat(r, a, &la_.IOpt{"n", m}, &la_.IOpt{"offset", m * i})
blas.ScalFloat(rti, a, &la_.IOpt{"n", m}, &la_.IOpt{"offset", m * i})
}
ind += m
ind2 += m * m
ind3 += m // !!NOT USED: ind3!!
}
//fmt.Printf("-- end of s:\nz=\n%v\nlmbda=\n%v\n", z.ConvertToString(), lmbda.ConvertToString())
return
}
/*
Applies Nesterov-Todd scaling or its inverse.
Computes
x := W*x (trans is false 'N', inverse = false 'N')
x := W^T*x (trans is true 'T', inverse = false 'N')
x := W^{-1}*x (trans is false 'N', inverse = true 'T')
x := W^{-T}*x (trans is true 'T', inverse = true 'T').
x is a dense float matrix.
W is a MatrixSet with entries:
- W['dnl']: positive vector
- W['dnli']: componentwise inverse of W['dnl']
- W['d']: positive vector
- W['di']: componentwise inverse of W['d']
- W['v']: lists of 2nd order cone vectors with unit hyperbolic norms
- W['beta']: list of positive numbers
- W['r']: list of square matrices
- W['rti']: list of square matrices. rti[k] is the inverse transpose
of r[k].
The 'dnl' and 'dnli' entries are optional, and only present when the
function is called from the nonlinear solver.
*/
func scale(x *matrix.FloatMatrix, W *FloatMatrixSet, trans, inverse bool) (err error) {
/*DEBUGGED*/
var wl []*matrix.FloatMatrix
var w *matrix.FloatMatrix
ind := 0
err = nil
// Scaling for nonlinear component xk is xk := dnl .* xk; inverse
// scaling is xk ./ dnl = dnli .* xk, where dnl = W['dnl'],
// dnli = W['dnli'].
if wl = W.At("dnl"); wl != nil {
if inverse {
w = W.At("dnli")[0]
} else {
w = W.At("dnl")[0]
}
for k := 0; k < x.Cols(); k++ {
err = blas.TbmvFloat(w, x, &la_.IOpt{"n", w.Rows()}, &la_.IOpt{"k", 0},
&la_.IOpt{"lda", 1}, &la_.IOpt{"offsetx", k * x.Rows()})
if err != nil {
return
}
}
ind += w.Rows()
}
// Scaling for linear 'l' component xk is xk := d .* xk; inverse
// scaling is xk ./ d = di .* xk, where d = W['d'], di = W['di'].
if inverse {
w = W.At("di")[0]
} else {
w = W.At("d")[0]
}
for k := 0; k < x.Cols(); k++ {
err = blas.TbmvFloat(w, x, &la_.IOpt{"n", w.Rows()}, &la_.IOpt{"k", 0},
&la_.IOpt{"lda", 1}, &la_.IOpt{"offsetx", k*x.Rows() + ind})
if err != nil {
return
}
}
ind += w.Rows()
// Scaling for 'q' component is
//
// xk := beta * (2*v*v' - J) * xk
// = beta * (2*v*(xk'*v)' - J*xk)
//
// where beta = W['beta'][k], v = W['v'][k], J = [1, 0; 0, -I].
//
//Inverse scaling is
//
// xk := 1/beta * (2*J*v*v'*J - J) * xk
// = 1/beta * (-J) * (2*v*((-J*xk)'*v)' + xk).
//wf := matrix.FloatZeros(x.Cols(), 1)
w = matrix.FloatZeros(x.Cols(), 1)
for k, v := range W.At("v") {
m := v.Rows()
if inverse {
blas.ScalFloat(x, -1.0, &la_.IOpt{"offset", ind}, &la_.IOpt{"inc", x.Rows()})
}
err = blas.GemvFloat(x, v, w, 1.0, 0.0, la_.OptTrans, &la_.IOpt{"m", m},
&la_.IOpt{"n", x.Cols()}, &la_.IOpt{"offsetA", ind},
&la_.IOpt{"lda", x.Rows()})
if err != nil {
return
}
err = blas.ScalFloat(x, -1.0, &la_.IOpt{"offset", ind}, &la_.IOpt{"inc", x.Rows()})
if err != nil {
return
}
err = blas.GerFloat(v, w, x, 2.0, &la_.IOpt{"m", m},
&la_.IOpt{"n", x.Cols()}, &la_.IOpt{"lda", x.Rows()},
&la_.IOpt{"offsetA", ind})
if err != nil {
return
}
var a float64
if inverse {
blas.ScalFloat(x, -1.0,
&la_.IOpt{"offset", ind}, &la_.IOpt{"inc", x.Rows()})
// a[i,j] := 1.0/W[i,j]
a = 1.0 / W.At("beta")[0].GetIndex(k)
} else {
a = W.At("beta")[0].GetIndex(k)
}
for i := 0; i < x.Cols(); i++ {
blas.ScalFloat(x, a, &la_.IOpt{"n", m}, &la_.IOpt{"offset", ind + i*x.Rows()})
}
ind += m
}
// Scaling for 's' component xk is
//
// xk := vec( r' * mat(xk) * r ) if trans = 'N'
// xk := vec( r * mat(xk) * r' ) if trans = 'T'.
//
// r is kth element of W['r'].
//
// Inverse scaling is
//
// xk := vec( rti * mat(xk) * rti' ) if trans = 'N'
// xk := vec( rti' * mat(xk) * rti ) if trans = 'T'.
//
// rti is kth element of W['rti'].
maxn := 0
for _, r := range W.At("r") {
if r.Rows() > maxn {
maxn = r.Rows()
}
}
a := matrix.FloatZeros(maxn, maxn)
for k, v := range W.At("r") {
t := trans
var r *matrix.FloatMatrix
if !inverse {
r = v
t = !trans
} else {
r = W.At("rti")[k]
}
n := r.Rows()
for i := 0; i < x.Cols(); i++ {
// scale diagonal of xk by 0.5
blas.ScalFloat(x, 0.5, &la_.IOpt{"offset", ind + i*x.Rows()},
&la_.IOpt{"inc", n + 1}, &la_.IOpt{"n", n})
// a = r*tril(x) (t is 'N') or a = tril(x)*r (t is 'T')
blas.Copy(r, a)
if !t {
err = blas.TrmmFloat(x, a, 1.0, la_.OptRight, &la_.IOpt{"m", n},
&la_.IOpt{"n", n}, &la_.IOpt{"lda", n}, &la_.IOpt{"ldb", n},
&la_.IOpt{"offsetA", ind + i*x.Rows()})
if err != nil {
return
}
// x := (r*a' + a*r') if t is 'N'
err = blas.Syr2kFloat(r, a, x, 1.0, 0.0, la_.OptNoTrans, &la_.IOpt{"n", n},
&la_.IOpt{"k", n}, &la_.IOpt{"ldb", n}, &la_.IOpt{"ldc", n},
&la_.IOpt{"offsetC", ind + i*x.Rows()})
if err != nil {
return
}
} else {
err = blas.TrmmFloat(x, a, 1.0, la_.OptLeft, &la_.IOpt{"m", n},
&la_.IOpt{"n", n}, &la_.IOpt{"lda", n}, &la_.IOpt{"ldb", n},
&la_.IOpt{"offsetA", ind + i*x.Rows()})
if err != nil {
return
}
// x := (r'*a + a'*r) if t is 'T'
err = blas.Syr2kFloat(r, a, x, 1.0, 0.0, la_.OptTrans, &la_.IOpt{"n", n},
&la_.IOpt{"k", n}, &la_.IOpt{"ldb", n}, &la_.IOpt{"ldc", n},
&la_.IOpt{"offsetC", ind + i*x.Rows()})
if err != nil {
return
}
}
}
ind += n * n
}
return
}
/*
Evaluates
x := H(lambda^{1/2}) * x (inverse is 'N')
x := H(lambda^{-1/2}) * x (inverse is 'I').
H is the Hessian of the logarithmic barrier.
*/
func scale2(lmbda, x *matrix.FloatMatrix, dims *DimensionSet, mnl int, inverse bool) (err error) {
err = nil
// For the nonlinear and 'l' blocks,
//
// xk := xk ./ l (inverse is 'N')
// xk := xk .* l (inverse is 'I')
//
// where l is lmbda[:mnl+dims['l']].
ind := mnl + dims.Sum("l")
if !inverse {
blas.TbsvFloat(lmbda, x, &la_.IOpt{"n", ind}, &la_.IOpt{"k", 0}, &la_.IOpt{"lda", 1})
} else {
blas.TbmvFloat(lmbda, x, &la_.IOpt{"n", ind}, &la_.IOpt{"k", 0}, &la_.IOpt{"lda", 1})
}
// For 'q' blocks, if inverse is 'N',
//
// xk := 1/a * [ l'*J*xk;
// xk[1:] - (xk[0] + l'*J*xk) / (l[0] + 1) * l[1:] ].
//
// If inverse is 'I',
//
// xk := a * [ l'*xk;
// xk[1:] + (xk[0] + l'*xk) / (l[0] + 1) * l[1:] ].
//
// a = sqrt(lambda_k' * J * lambda_k), l = lambda_k / a.
for _, m := range dims.At("q") {
var lx, a, c, x0 float64
a = jnrm2(lmbda, m, ind) //&la_.IOpt{"n", m}, &la_.IOpt{"offset", ind})
if !inverse {
lx = jdot(lmbda, x, m, ind, ind) //&la_.IOpt{"n", m}, &la_.IOpt{"offsetx", ind},
//&la_.IOpt{"offsety", ind})
lx /= a
} else {
lx = blas.DotFloat(lmbda, x, &la_.IOpt{"n", m}, &la_.IOpt{"offsetx", ind},
&la_.IOpt{"offsety", ind})
lx /= a
}
x0 = x.GetIndex(ind)
x.SetIndex(ind, lx)
c = (lx + x0) / (lmbda.GetIndex(ind)/a + 1.0) / a
if !inverse {
c *= -1.0
}
blas.AxpyFloat(lmbda, x, c, &la_.IOpt{"n", m - 1}, &la_.IOpt{"offsetx", ind + 1},
&la_.IOpt{"offsety", ind + 1})
if !inverse {
a = 1.0 / a
}
blas.ScalFloat(x, a, &la_.IOpt{"offset", ind}, &la_.IOpt{"n", m})
ind += m
}
// For the 's' blocks, if inverse is 'N',
//
// xk := vec( diag(l)^{-1/2} * mat(xk) * diag(k)^{-1/2}).
//
// If inverse is true,
//
// xk := vec( diag(l)^{1/2} * mat(xk) * diag(k)^{1/2}).
//
// where l is kth block of lambda.
//
// We scale upper and lower triangular part of mat(xk) because the
// inverse operation will be applied to nonsymmetric matrices.
ind2 := ind
sdims := dims.At("s")
for k := 0; k < len(sdims); k++ {
m := sdims[k]
scaleF := func(v, x float64) float64 {
return math.Sqrt(v) * math.Sqrt(x)
}
for j := 0; j < m; j++ {
c := matrix.FloatVector(lmbda.FloatArray()[ind2 : ind2+m])
c.ApplyConst(c, scaleF, lmbda.GetIndex(ind2+j))
if !inverse {
blas.Tbsv(c, x, &la_.IOpt{"n", m}, &la_.IOpt{"k", 0}, &la_.IOpt{"lda", 1},
&la_.IOpt{"offsetx", ind + j*m})
} else {
blas.Tbmv(c, x, &la_.IOpt{"n", m}, &la_.IOpt{"k", 0}, &la_.IOpt{"lda", 1},
&la_.IOpt{"offsetx", ind + j*m})
}
}
ind += m * m
ind2 += m
}
return
}
