From 67fc9270709b908ee79cbb65c1d28745722a089d Mon Sep 17 00:00:00 2001
From: Michael Baudin
Date: Fri, 26 Sep 2008 15:19:40 +0000
Subject: [PATCH] Added references. Clarified link with quapro. Added memory
requirements.

.../modules/optimization/help/en_US/qp_solve.xml  64 +++++++++++++
scilab/modules/optimization/help/en_US/qpsolve.xml  69 +++++++++++++++++
2 files changed, 108 insertions(+), 25 deletions()
diff git a/scilab/modules/optimization/help/en_US/qp_solve.xml b/scilab/modules/optimization/help/en_US/qp_solve.xml
index fece078..00b427e 100644
 a/scilab/modules/optimization/help/en_US/qp_solve.xml
+++ b/scilab/modules/optimization/help/en_US/qp_solve.xml
@@ 48,8 +48,7 @@
real matrix (dimension (me + md) x n).
 This matrix may be stored as either a dense matrix either a sparse
 one.
+ This matrix may be dense or sparse.
@@ 57,7 +56,7 @@
b
 RHS column vector (dimension (me +
+ RHS column vector (dimension m=(me +
md))
@@ 92,7 +91,7 @@
iter
 . 2x1 vector, first component gives the number of "main"
+ 2x1 vector, first component gives the number of "main"
iterations, the second one says how many constraints were deleted
after they became active.
@@ 114,14 +113,9 @@
 This function requires Q to be positive definite,
 if it is not the case, one may use the contributed quapro toolbox.

 This is a builtin associated with the Berwin A. Turlach code. Use
 preferabily the qpsolve function built on it
 which handles variable boundaries and match the quapro problem definition
+ This function requires Q to be symmetric positive
+ definite. If this hypothesis is not satisfied, one may use the contributed
+ quapro toolbox.
@@ 165,6 +159,7 @@ me=3;
optimqld
+
qpsolve
@@ 173,6 +168,19 @@ me=3;
+ Memory requirements
+
+ Let r be
+
+ r=min(m,n)
+
+ Then the memory required by qp_solve during the computations is
+
+
+ 2*n+r*(r+5)/2 + 2*m +1
+
+
+ Authors
@@ 180,7 +188,7 @@ me=3;
S. Steer
 , INRIA (Scilab interface)
+ INRIA (Scilab interface)
@@ 188,14 +196,40 @@ me=3;
Berwin A. Turlach
 ,School of Mathematics and Statistics (M019), The University
 of Western Australia, Crawley, AUSTRALIA (solver code)
+ School of Mathematics and Statistics (M019), The University of
+ Western Australia, Crawley, AUSTRALIA (solver code)
+ References
+
+
+
+ Goldfarb, D. and Idnani, A. (1982). "Dual and PrimalDual
+ Methods for Solving Strictly Convex Quadratic Programs", in J.P.
+ Hennart (ed.), Numerical Analysis, Proceedings, Cocoyoc, Mexico 1981,
+ Vol. 909 of Lecture Notes in Mathematics, SpringerVerlag, Berlin, pp.
+ 226239.
+
+
+
+ Goldfarb, D. and Idnani, A. (1983). "A numerically stable dual
+ method for solving strictly convex quadratic programs", Mathematical
+ Programming 27: 133.
+
+
+
+ QuadProg (Quadratic Programming Routines), Berwin A
+ Turlach,http://www.maths.uwa.edu.au/~berwin/software/quadprog.html
+
+
+
+
+ Used Functionsqpgen2.f and >qpgen1.f (also named QP.solve.f) developped by
diff git a/scilab/modules/optimization/help/en_US/qpsolve.xml b/scilab/modules/optimization/help/en_US/qpsolve.xml
index 9ee5c15..24f0a19 100644
 a/scilab/modules/optimization/help/en_US/qpsolve.xml
+++ b/scilab/modules/optimization/help/en_US/qpsolve.xml
@@ 48,8 +48,7 @@
real matrix (dimension (me + md) x n).
 This matrix may be stored as either a dense matrix either a sparse
 one.
+ This matrix may be dense or sparse.
@@ 57,7 +56,7 @@
b
 RHS column vector (dimension (me +
+ RHS column vector (dimension m=(me +
md))
@@ 137,9 +136,17 @@
 This function requires Q to be positive definite,
 if it is not the case, one may use the quapro function.
+ This function requires Q to be symmetric positive
+ definite. If that hypothesis is not satisfied, one may use the quapro function, which is provided in the Scilab
+ quapro toolbox.
+
+ The qpsolve solver is implemented as a Scilab script, which calls
+ the compiled qp_solve primitive. It is provided as a facility, in order to
+ be a direct replacement for the former quapro solver : indeed, the qpsolve
+ solver has been designed so that it provides the same interface, that is,
+ the same input/output arguments. But the x0 and imp input arguments are
+ available in quapro, but not in qpsolve.
@@ 176,11 +183,27 @@ me=3;
optim
+
qp_solve
+
qld
+
The contributed toolbox "quapro" may also be of interest, in
 particular for singular Q.
+ particular for singular Q.
+
+
+
+ Memory requirements
+
+ Let r be
+
+ r=min(m,n)
+
+ Then the memory required by qpsolve during the computations
+ is
+
+ 2*n+r*(r+5)/2 + 2*m +1
@@ 191,7 +214,7 @@ me=3;
S. Steer
 , INRIA (Scilab interface)
+ INRIA (Scilab interface)
@@ 199,14 +222,40 @@ me=3;
Berwin A. Turlach
 ,School of Mathematics and Statistics (M019), The University
 of Western Australia, Crawley, AUSTRALIA (solver code)
+ School of Mathematics and Statistics (M019), The University of
+ Western Australia, Crawley, AUSTRALIA (solver code)
+ References
+
+
+
+ Goldfarb, D. and Idnani, A. (1982). "Dual and PrimalDual
+ Methods for Solving Strictly Convex Quadratic Programs", in J.P.
+ Hennart (ed.), Numerical Analysis, Proceedings, Cocoyoc, Mexico 1981,
+ Vol. 909 of Lecture Notes in Mathematics, SpringerVerlag, Berlin, pp.
+ 226239.
+
+
+
+ Goldfarb, D. and Idnani, A. (1983). "A numerically stable dual
+ method for solving strictly convex quadratic programs", Mathematical
+ Programming 27: 133.
+
+
+
+ QuadProg (Quadratic Programming Routines), Berwin A
+ Turlach,http://www.maths.uwa.edu.au/~berwin/software/quadprog.html
+
+
+
+
+ Used Functionsqpgen1.f (also named QP.solve.f) developped by Berwin A. Turlach

1.7.9.5