PDF version (NAG web site
, 64bit version, 64bit version)
NAG Toolbox: nag_opt_lsq_lincon_solve (e04nc)
Purpose
nag_opt_lsq_lincon_solve (e04nc) solves linearly constrained linear least squares problems and convex quadratic programming problems. It is not intended for large sparse problems.
Syntax
[
istate,
kx,
x,
a,
b,
iter,
obj,
clamda,
lwsav,
iwsav,
rwsav,
ifail] = e04nc(
c,
bl,
bu,
cvec,
istate,
kx,
x,
a,
b,
lwsav,
iwsav,
rwsav, 'm',
m, 'n',
n, 'nclin',
nclin)
[
istate,
kx,
x,
a,
b,
iter,
obj,
clamda,
lwsav,
iwsav,
rwsav,
ifail] = nag_opt_lsq_lincon_solve(
c,
bl,
bu,
cvec,
istate,
kx,
x,
a,
b,
lwsav,
iwsav,
rwsav, 'm',
m, 'n',
n, 'nclin',
nclin)
Before calling
nag_opt_lsq_lincon_solve (e04nc), or
the option setting function
(e04ne),
nag_opt_init (e04wb) must be called.
Description
nag_opt_lsq_lincon_solve (e04nc) is designed to solve a class of quadratic programming problems of the following general form:
where
c${\mathbf{c}}$ is an
n_{L}${n}_{L}$ by
n$n$ matrix and the objective function
F(x)$F\left(x\right)$ may be specified in a variety of ways depending upon the particular problem to be solved. The available forms for
F(x)$F\left(x\right)$ are listed in
Table 1, in which the prefixes FP, LP, QP and LS stand for ‘feasible point’, ‘linear programming’, ‘quadratic programming’ and ‘least squares’ respectively,
c$c$ is an
n$n$element vector,
b$b$ is an
m$m$ element vector and
‖z‖$\Vert z\Vert $ denotes the Euclidean length of
z$z$.
Problem type 
F(x)$F\left(x\right)$ 
Matrix A$A$ 
FP 
None 
Not applicable 
LP 
c^{T}x${c}^{\mathrm{T}}x$ 
Not applicable 
QP1 
c^{T}x + (1/2)x^{T}Ax$\phantom{{c}^{\mathrm{T}}x+}\frac{1}{2}{x}^{\mathrm{T}}Ax$ 
n$n$ by n$n$ symmetric positive semidefinite 
QP2 
c^{T}x + (1/2)x^{T}Ax${c}^{\mathrm{T}}x+\frac{1}{2}{x}^{\mathrm{T}}Ax$ 
n$n$ by n$n$ symmetric positive semidefinite 
QP3 
c^{T}x + (1/2)x^{T}A^{T}Ax$\phantom{{c}^{\mathrm{T}}x+}\frac{1}{2}{x}^{\mathrm{T}}{A}^{\mathrm{T}}Ax$ 
m$m$ by n$n$ upper trapezoidal 
QP4 
c^{T}x + (1/2)x^{T}A^{T}Ax${c}^{\mathrm{T}}x+\frac{1}{2}{x}^{\mathrm{T}}{A}^{\mathrm{T}}Ax$ 
m$m$ by n$n$ upper trapezoidal 
LS1 
c^{T}x + (1/2)‖b − Ax‖^{2}$\phantom{{c}^{\mathrm{T}}x+}\frac{1}{2}{\Vert bAx\Vert}^{2}$ 
m$m$ by n$n$ 
LS2 
c^{T}x + (1/2)‖b − Ax‖^{2}${c}^{\mathrm{T}}x+\frac{1}{2}{\Vert bAx\Vert}^{2}$ 
m$m$ by n$n$ 
LS3 
c^{T}x + (1/2)‖b − Ax‖^{2}$\phantom{{c}^{\mathrm{T}}x+}\frac{1}{2}{\Vert bAx\Vert}^{2}$ 
m$m$ by n$n$ upper trapezoidal 
LS4 
c^{T}x + (1/2)‖b − Ax‖^{2}${c}^{\mathrm{T}}x+\frac{1}{2}{\Vert bAx\Vert}^{2}$ 
m$m$ by n$n$ upper trapezoidal 
Table 1
In the standard LS problem
F(x)$F\left(x\right)$ will usually have the form LS1, and in the standard convex QP problem
F(x)$F\left(x\right)$ will usually have the form QP2. The default problem type is LS1 and other objective functions are selected by using the optional parameter
Problem Type.
When A$A$ is upper trapezoidal it will usually be the case that m = n$m=n$, so that A$A$ is upper triangular, but full generality has been allowed for in the specification of the problem. The upper trapezoidal form is intended for cases where a previous factorization, such as a QR$QR$ factorization, has been performed.
The constraints involving
c${\mathbf{c}}$ are called the
general constraints. Note that upper and lower bounds are specified for all the variables and for all the general constraints. An equality constraint can be specified by setting
l_{i} = u_{i}${l}_{i}={u}_{i}$. If certain bounds are not present, the associated elements of
l$l$ or
u$u$ can be set to special values that will be treated as
− ∞$\infty $ or
+ ∞$+\infty $. (See the description of the optional parameter
Infinite Bound Size.)
The defining feature of a quadratic function F(x)$F\left(x\right)$ is that the secondderivative matrix H$H$ (the Hessian matrix) is constant. For the LP case H = 0$H=0$; for QP1 and QP2, H = A$H=A$; for QP3 and QP4, H = A^{T}A$H={A}^{\mathrm{T}}A$ and for LS1 (the default), LS2, LS3 and LS4, H = A^{T}A$H={A}^{\mathrm{T}}A$.
Problems of type QP3 and QP4 for which A$A$ is not in upper trapezoidal form should be solved as types LS1 and LS2 respectively, with b = 0$b=0$.
For problems of type LS, we refer to A$A$ as the least squares matrix, or the matrix of observations and to b$b$ as the vector of observations.
You must supply an initial estimate of the solution.
If H$H$ is nonsingular then nag_opt_lsq_lincon_solve (e04nc) will obtain the unique (global) minimum. If H$H$ is singular then the solution may still be a global minimum if all active constraints have nonzero Lagrange multipliers. Otherwise the solution obtained will be either a weak minimum (i.e., with a unique optimal objective value, but an infinite set of optimal x$x$), or else the objective function is unbounded below in the feasible region. The last case can only occur when F(x)$F\left(x\right)$ contains an explicit linear term (as in problems LP, QP2, QP4, LS2 and LS4).
The method used by
nag_opt_lsq_lincon_solve (e04nc) is described in detail in
Section [Algorithmic Details].
References
Gill P E, Hammarling S, Murray W, Saunders M A and Wright M H (1986) Users' guide for LSSOL (Version 1.0) Report SOL 861 Department of Operations Research, Stanford University
Gill P E, Murray W, Saunders M A and Wright M H (1984) Procedures for optimization problems with a mixture of bounds and general linear constraints ACM Trans. Math. Software 10 282–298
Gill P E, Murray W and Wright M H (1981) Practical Optimization Academic Press
Stoer J (1971) On the numerical solution of constrained least squares problems SIAM J. Numer. Anal. 8 382–411
Parameters
Compulsory Input Parameters
 1:
c(ldc, : $:$) – double array

The first dimension of the array
c must be at least
max (1,nclin)$\mathrm{max}\phantom{\rule{0.125em}{0ex}}(1,{\mathbf{nclin}})$The second dimension of the array must be at least
n${\mathbf{n}}$ if
nclin > 0${\mathbf{nclin}}>0$, and at least
1$1$ otherwise
The
i$\mathit{i}$th row of
c must contain the coefficients of the
i$\mathit{i}$th general constraint, for
i = 1,2, … ,nclin$\mathit{i}=1,2,\dots ,{\mathbf{nclin}}$.
If
nclin = 0${\mathbf{nclin}}=0$,
c is not referenced.
 2:
bl(n + nclin${\mathbf{n}}+{\mathbf{nclin}}$) – double array
 3:
bu(n + nclin${\mathbf{n}}+{\mathbf{nclin}}$) – double array
bl must contain the lower bounds and
bu the upper bounds, for all the constraints, in the following order. The first
n$n$ elements of each array must contain the bounds on the variables, and the next
n_{L}${n}_{L}$ elements must contain the bounds for the general linear constraints (if any). To specify a nonexistent lower bound (i.e.,
l_{j} = − ∞${l}_{j}=\infty $), set
bl(j) ≤ − bigbnd${\mathbf{bl}}\left(j\right)\le \mathit{bigbnd}$, and to specify a nonexistent upper bound (i.e.,
u_{j} = + ∞${u}_{j}=+\infty $), set
bu(j) ≥ bigbnd${\mathbf{bu}}\left(j\right)\ge \mathit{bigbnd}$; the default value of
bigbnd$\mathit{bigbnd}$ is
10^{20}${10}^{20}$, but this may be changed by the optional parameter
Infinite Bound Size. To specify the
j$j$th constraint as an equality, set
bu(j) = bl(j) = β${\mathbf{bu}}\left(j\right)={\mathbf{bl}}\left(j\right)=\beta $, say, where
β < bigbnd$\left\beta \right<\mathit{bigbnd}$.
Constraints:
 bl(j) ≤ bu(j)${\mathbf{bl}}\left(\mathit{j}\right)\le {\mathbf{bu}}\left(\mathit{j}\right)$, for j = 1,2, … ,n + nclin$\mathit{j}=1,2,\dots ,{\mathbf{n}}+{\mathbf{nclin}}$;
 if bl(j) = bu(j) = β${\mathbf{bl}}\left(j\right)={\mathbf{bu}}\left(j\right)=\beta $, β < bigbnd$\left\beta \right<\mathit{bigbnd}$.
 4:
cvec( : $:$) – double array

Note: the dimension of the array
cvec
must be at least
n${\mathbf{n}}$ if the problem is of type LP, QP2, QP4, LS2 or LS4, and at least
1$1$ otherwise.
The coefficients of the explicit linear term of the objective function.
If the problem is of type FP, QP1, QP3, LS1 (the default) or LS3,
cvec is not referenced.
 5:
istate(n + nclin${\mathbf{n}}+{\mathbf{nclin}}$) – int64int32nag_int array
Need not be set if the (default) optional parameter
Cold Start is used.
If the optional parameter
Warm Start has been chosen,
istate specifies the desired status of the constraints at the start of the feasibility phase. More precisely, the first
n$n$ elements of
istate refer to the upper and lower bounds on the variables, and the next
n_{L}${n}_{L}$ elements refer to the general linear constraints (if any). Possible values for
istate(j)${\mathbf{istate}}\left(j\right)$ are as follows:
istate(j)${\mathbf{istate}}\left(j\right)$  Meaning 
0  The constraint should not be in the initial working set. 
1  The constraint should be in the initial working set at its lower bound. 
2  The constraint should be in the initial working set at its upper bound. 
3  The constraint should be in the initial working set as an equality. This value must not be specified unless bl(j) = bu(j)${\mathbf{bl}}\left(j\right)={\mathbf{bu}}\left(j\right)$. 
The values
− 2$2$,
− 1$1$ and
4$4$ are also acceptable but will be reset to zero by the function. If
nag_opt_lsq_lincon_solve (e04nc) has been called previously with the same values of
n and
nclin,
istate already contains satisfactory information. (See also the description of the optional parameter
Warm Start.) The function also adjusts (if necessary) the values supplied in
x to be consistent with
istate.
Constraint:
− 2 ≤ istate(j) ≤ 4$2\le {\mathbf{istate}}\left(\mathit{j}\right)\le 4$, for
j = 1,2, … ,n + nclin$\mathit{j}=1,2,\dots ,{\mathbf{n}}+{\mathbf{nclin}}$.
 6:
kx(n) – int64int32nag_int array
n, the dimension of the array, must satisfy the constraint
n > 0${\mathbf{n}}>0$.
Need not be initialized for problems of type FP, LP, QP1, QP2, LS1 (the default) or LS2.
For problems QP3, QP4, LS3 or LS4,
kx must specify the order of the columns of the matrix
A$A$ with respect to the ordering of
x. Thus if column
j$j$ of
A$A$ is the column associated with the variable
x_{i}${x}_{i}$ then
kx(j) = i${\mathbf{kx}}\left(j\right)=i$.
Constraints:
 1 ≤ kx(i) ≤ n$1\le {\mathbf{kx}}\left(\mathit{i}\right)\le {\mathbf{n}}$, for i = 1,2, … ,n$\mathit{i}=1,2,\dots ,{\mathbf{n}}$;
 if i ≠ j$i\ne j$, kx(i) ≠ kx(j)${\mathbf{kx}}\left(i\right)\ne {\mathbf{kx}}\left(j\right)$.
 7:
x(n) – double array

n, the dimension of the array, must satisfy the constraint
n > 0${\mathbf{n}}>0$.
An initial estimate of the solution.
Note: that it may be best to avoid the choice
x = 0.0${\mathbf{x}}=0.0$.
 8:
a(lda, : $:$) – double array
The first dimension of the array
a must be at least
max (1,m)$\mathrm{max}\phantom{\rule{0.125em}{0ex}}(1,{\mathbf{m}})$The second dimension of the array must be at least
n${\mathbf{n}}$ if the problem is of type QP1, QP2, QP3, QP4, LS1 (the default), LS2, LS3 or LS4, and at least
1$1$ otherwise
The array
a must contain the matrix
A$A$ as specified in
Table 1 (see
Section [Description]).
If the problem is of type QP1 or QP2, the first
m$m$ rows and columns of
a must contain the leading
m$m$ by
m$m$ rows and columns of the symmetric Hessian matrix. Only the diagonal and upper triangular elements of the leading
m$m$ rows and columns of
a are referenced. The remaining elements are assumed to be zero and need not be assigned.
For problems QP3, QP4, LS3 or LS4, the first
m$m$ rows of
a must contain an
m$m$ by
n$n$ upper trapezoidal factor of either the Hessian matrix or the least squares matrix, ordered according to the
kx array. The factor need not be of full rank, i.e., some of the diagonals may be zero. However, as a general rule, the larger the dimension of the leading nonsingular submatrix of
A$A$, the fewer iterations will be required. Elements outside the upper triangular part of the first
m$m$ rows of
a are assumed to be zero and need not be assigned.
If a constrained least squares problem contains a very large number of observations, storage limitations may prevent storage of the entire least squares matrix. In such cases, you should transform the original A$A$ into a triangular matrix before the call to nag_opt_lsq_lincon_solve (e04nc) and solve the problem as type LS3 or LS4.
 9:
b( : $:$) – double array
Note: the dimension of the array
b
must be at least
m${\mathbf{m}}$ if the problem is of type LS1 (the default), LS2, LS3 or LS4, and at least
1$1$ otherwise.
The m$m$ elements of the vector of observations.
 10:
lwsav(120$120$) – logical array
 11:
iwsav(610$610$) – int64int32nag_int array
 12:
rwsav(475$475$) – double array
The arrays
lwsav,
iwsav and
rwsav must not be altered between calls to any of the functions
nag_opt_lsq_lincon_solve (e04nc),
(e04nd) or
(e04ne).
Optional Input Parameters
 1:
m – int64int32nag_int scalar
Default:
The dimension of the array
a.
m$m$, the number of rows in the matrix
A$A$. If the problem is specified as type FP or LP,
m is not referenced and is assumed to be zero.
If the problem is of type QP,
m will usually be
n$n$, the number of variables. However, a value of
m less than
n$n$ is appropriate for QP3 or QP4 if
A$A$ is an upper trapezoidal matrix with
m$m$ rows. Similarly,
m may be used to define the dimension of a leading block of nonzeros in the Hessian matrices of QP1 or QP2, in which case the last
(n − m)$(nm)$ rows and columns of
a are assumed to be zero. In the QP case,
m$m$ should not be greater than
n$n$; if it is, the last
(m − n)$(mn)$ rows of
A$A$ are ignored.
If the problem is of type LS1 (the default) or specified as type LS2, LS3 or LS4,
m is also the dimension of the array
b. Note that all possibilities (
m < n$m<n$,
m = n$m=n$ and
m > n$m>n$) are allowed in this case.
Constraint:
m > 0${\mathbf{m}}>0$ if the problem is not of type FP or LP.
 2:
n – int64int32nag_int scalar
Default:
The dimension of the arrays
kx,
x. (An error is raised if these dimensions are not equal.)
n$n$, the number of variables.
Constraint:
n > 0${\mathbf{n}}>0$.
 3:
nclin – int64int32nag_int scalar
Default:
The dimension of the array
c.
n_{L}${n}_{L}$, the number of general linear constraints.
Constraint:
nclin ≥ 0${\mathbf{nclin}}\ge 0$.
Input Parameters Omitted from the MATLAB Interface
 ldc lda iwork liwork work lwork
Output Parameters
 1:
istate(n + nclin${\mathbf{n}}+{\mathbf{nclin}}$) – int64int32nag_int array
The status of the constraints in the working set at the point returned in
x. The significance of each possible value of
istate(j)${\mathbf{istate}}\left(j\right)$ is as follows:
istate(j)${\mathbf{istate}}\left(j\right)$  Meaning 
− 2$2$  The constraint violates its lower bound by more than the feasibility tolerance. 
− 1$1$  The constraint violates its upper bound by more than the feasibility tolerance. 
− 0$\phantom{}0$  The constraint is satisfied to within the feasibility tolerance, but is not in the working set. 
− 1$\phantom{}1$  This inequality constraint is included in the working set at its lower bound. 
− 2$\phantom{}2$  This inequality constraint is included in the working set at its upper bound. 
− 3$\phantom{}3$  The constraint is included in the working set as an equality. This value of istate can occur only when bl(j) = bu(j)${\mathbf{bl}}\left(j\right)={\mathbf{bu}}\left(j\right)$. 
− 4$\phantom{}4$  This corresponds to optimality being declared with x(j)${\mathbf{x}}\left(j\right)$ being temporarily fixed at its current value. 
 2:
kx(n) – int64int32nag_int array
Defines the order of the columns of
a with respect to the ordering of
x, as described above.
 3:
x(n) – double array
The point at which
nag_opt_lsq_lincon_solve (e04nc) terminated. If
ifail = 0${\mathbf{ifail}}={\mathbf{0}}$,
1${\mathbf{1}}$ or
4${\mathbf{4}}$,
x contains an estimate of the solution.
 4:
a(lda, : $:$) – double array
The first dimension of the array
a will be
max (1,m)$\mathrm{max}\phantom{\rule{0.125em}{0ex}}(1,{\mathbf{m}})$The second dimension of the array will be
n${\mathbf{n}}$ if the problem is of type QP1, QP2, QP3, QP4, LS1 (the default), LS2, LS3 or LS4, and at least
1$1$ otherwise
lda ≥ max (1,m)$\mathit{lda}\ge \mathrm{max}\phantom{\rule{0.125em}{0ex}}(1,{\mathbf{m}})$.
If
Hessian = NO${\mathbf{Hessian}}=\mathrm{NO}$ and the problem is of type LS or QP,
a contains the upper triangular Cholesky factor
R$R$ of
(8) (see
Section [Main Iteration]), with columns ordered as indicated by
kx. If
Hessian = YES${\mathbf{Hessian}}=\mathrm{YES}$ and the problem is of type LS or QP,
a contains the upper triangular Cholesky factor
R$R$ of the Hessian matrix
H$H$, with columns ordered as indicated by
kx. In either case
R$R$ may be used to obtain the variancecovariance matrix or to recover the upper triangular factor of the original least squares matrix.
If the problem is of type FP or LP,
a is not referenced.
 5:
b( : $:$) – double array
Note: the dimension of the array
b
must be at least
m${\mathbf{m}}$ if the problem is of type LS1 (the default), LS2, LS3 or LS4, and at least
1$1$ otherwise.
The transformed residual vector of equation
(10) (see
Section [Main Iteration]).
If the problem is of type FP, LP, QP1, QP2, QP3 or QP4,
b is not referenced.
 6:
iter – int64int32nag_int scalar
The total number of iterations performed.
 7:
obj – double scalar
The value of the objective function at
x$x$ if
x$x$ is feasible, or the sum of infeasibiliites at
x$x$ otherwise. If the problem is of type FP and
x$x$ is feasible,
obj is set to zero.
 8:
clamda(n + nclin${\mathbf{n}}+{\mathbf{nclin}}$) – double array
The values of the Lagrange multipliers for each constraint with respect to the current working set. The first
n$n$ elements contain the multipliers for the bound constraints on the variables, and the next
n_{L}${n}_{L}$ elements contain the multipliers for the general linear constraints (if any). If
istate(j) = 0${\mathbf{istate}}\left(j\right)=0$ (i.e., constraint
j$j$ is not in the working set),
clamda(j)${\mathbf{clamda}}\left(j\right)$ is zero. If
x$x$ is optimal,
clamda(j)${\mathbf{clamda}}\left(j\right)$ should be nonnegative if
istate(j) = 1${\mathbf{istate}}\left(j\right)=1$, nonpositive if
istate(j) = 2${\mathbf{istate}}\left(j\right)=2$ and zero if
istate(j) = 4${\mathbf{istate}}\left(j\right)=4$.
 9:
lwsav(120$120$) – logical array
 10:
iwsav(610$610$) – int64int32nag_int array
 11:
rwsav(475$475$) – double array
 12:
ifail – int64int32nag_int scalar
ifail = 0${\mathrm{ifail}}={\mathbf{0}}$ unless the function detects an error (see
[Error Indicators and Warnings]).
nag_opt_lsq_lincon_solve (e04nc) returns with
ifail = 0${\mathbf{ifail}}={\mathbf{0}}$ if
x$x$ is a strong local minimizer, i.e., the projected gradient (
Norm Gz; see
Section [Printed output]) is negligible, the Lagrange multipliers (
Lagr Mult; see
Section [Definition of Search Direction]) are optimal and
R_{Z}${R}_{Z}$ (see
Section [Main Iteration]) is nonsingular.
Error Indicators and Warnings
Note: nag_opt_lsq_lincon_solve (e04nc) may return useful information for one or more of the following detected errors or warnings.
Errors or warnings detected by the function:
Cases prefixed with W are classified as warnings and
do not generate an error of type NAG:error_n. See nag_issue_warnings.
 W ifail = 1${\mathbf{ifail}}=1$
x is a weak local minimum, (i.e., the projected gradient is negligible, the Lagrange multipliers are optimal, but either
R_{Z}${R}_{Z}$ (see
Section [Main Iteration]) is singular, or there is a small multiplier). This means that
x$x$ is not unique.
 W ifail = 2${\mathbf{ifail}}=2$
The solution appears to be unbounded. This value of
ifail implies that a step as large as
Infinite Bound Size (
default value = 10^{20}$\text{default value}={10}^{20}$) would have to be taken in order to continue the algorithm. This situation can occur only when
A$A$ is singular, there is an explicit linear term, and at least one variable has no upper or lower bound.
 W ifail = 3${\mathbf{ifail}}=3$
No feasible point was found, i.e., it was not possible to satisfy all the constraints to within the feasibility tolerance. In this case, the constraint violations at the final
x$x$ will reveal a value of the tolerance for which a feasible point will exist – for example, when the feasibility tolerance for each violated constraint exceeds its
Slack (see
Section [Printed output]) at the final point. The modified problem (with an altered feasibility tolerance) may then be solved using a
Warm Start. You should check that there are no constraint redundancies. If the data for the constraints are accurate only to the absolute precision
σ$\sigma $, you should ensure that the value of the optional parameter
Feasibility Tolerance (
default value = sqrt(ε)$\text{default value}=\sqrt{\epsilon}$, where
ε$\epsilon $ is the
machine precision) is
greater than
σ$\sigma $. For example, if all elements of
c${\mathbf{c}}$ are of order unity and are accurate only to three decimal places, the
Feasibility Tolerance should be at least
10^{ − 3}${10}^{3}$.
 ifail = 4${\mathbf{ifail}}=4$
The limiting number of iterations (determined by the optional parameters
Feasibility Phase Iteration Limit (
default value = max (50,5(n + n_{L}))$\text{default value}=\mathrm{max}\phantom{\rule{0.125em}{0ex}}(50,5(n+{n}_{L}))$) and
Optimality Phase Iteration Limit (
default value = max (50,5(n + n_{L}))$\text{default value}=\mathrm{max}\phantom{\rule{0.125em}{0ex}}(50,5(n+{n}_{L}))$)) was reached before normal termination occurred. If the method appears to be making progress (e.g., the objective function is being satisfactorily reduced), either increase the iterations limit and rerun
nag_opt_lsq_lincon_solve (e04nc) or, alternatively, rerun
nag_opt_lsq_lincon_solve (e04nc) using the
Warm Start facility to specify the initial working set. If the iteration limit is already large, but some of the constraints could be nearly linearly dependent, check the monitoring information (see
Section [Description of Monitoring Information]) for a repeated pattern of constraints entering and leaving the working set. (Neardependencies are often indicated by wide variations in size in the diagonal elements of the matrix
T$T$ (see
Section [Definition of Search Direction]), which will be printed if
Print Level ≥ 30${\mathbf{Print\; Level}}\ge 30$ (
default value = 10$\text{default value}=10$). In this case, the algorithm could be cycling (see the comments for
ifail = 5${\mathbf{ifail}}={\mathbf{5}}$).
 ifail = 5${\mathbf{ifail}}=5$

The algorithm could be cycling, since a total of
50$50$ changes were made to the working set without altering
x$x$. You should check the monitoring information (see
Section [Description of Monitoring Information]) for a repeated pattern of constraint deletions and additions.
If a sequence of constraint changes is being repeated, the iterates are probably cycling. (
nag_opt_lsq_lincon_solve (e04nc) does not contain a method that is guaranteed to avoid cycling; such a method would be combinatorial in nature.) Cycling may occur in two circumstances: at a constrained stationary point where there are some small or zero Lagrange multipliers; or at a point (usually a vertex) where the constraints that are satisfied exactly are nearly linearly dependent. In the latter case, you have the option of identifying the offending dependent constraints and removing them from the problem, or restarting the run with a larger value of the optional parameter
Feasibility Tolerance (
default value = sqrt(ε)$\text{default value}=\sqrt{\epsilon}$, where
ε$\epsilon $ is the
machine precision). If
nag_opt_lsq_lincon_solve (e04nc) terminates with
ifail = 5${\mathbf{ifail}}={\mathbf{5}}$, but no suspicious pattern of constraint changes can be observed, it may be worthwhile to restart with the final
x$x$ (with or without the
Warm Start option).
Note: that this error exit may also occur if a poor starting point
x is supplied (for example,
x = 0.0${\mathbf{x}}=0.0$). You are advised to try a nonzero starting point.
 ifail = 6${\mathbf{ifail}}=6$

An input parameter is invalid.
 ifail = 7${\mathbf{ifail}}=7$

The problem to be solved is of type QP1 or QP2, but the Hessian matrix supplied in
a is not positive semidefinite.
 Overflow$\mathbf{\text{Overflow}}$

If the printed output before the overflow error contains a warning about serious illconditioning in the working set when adding the
j$j$th constraint, it may be possible to avoid the difficulty by increasing the magnitude of the
Feasibility Tolerance (
default value = sqrt(ε)$\text{default value}=\sqrt{\epsilon}$, where
ε$\epsilon $ is the
machine precision) and rerunning the program. If the message recurs even after this change, the offending linearly dependent constraint (with index ‘
j$j$’) must be removed from the problem.
Accuracy
nag_opt_lsq_lincon_solve (e04nc) implements a numerically stable active set strategy and returns solutions that are as accurate as the condition of the problem warrants on the machine.
Further Comments
This section contains some comments on scaling and a description of the printed output.
Scaling
Sensible scaling of the problem is likely to reduce the number of iterations required and make the problem less sensitive to perturbations in the data, thus improving the condition of the problem. In the absence of better information it is usually sensible to make the Euclidean lengths of each constraint of comparable magnitude. See the
E04 Chapter Introduction and
Gill et al. (1981) for further information and advice.
Description of the Printed Output
This section describes the intermediate printout and final printout produced by
nag_opt_lsq_lincon_solve (e04nc). The intermediate printout is a subset of the monitoring information produced by the function at every iteration (see
Section [Description of Monitoring Information]). You can control the level of printed output (see the description of the optional parameter
Print Level).
Note that the intermediate printout and final printout are produced only if
Print Level ≥ 10${\mathbf{Print\; Level}}\ge 10$ (the default for
nag_opt_lsq_lincon_solve (e04nc), by default no output is produced by
nag_opt_lsq_lincon_solve (e04nc)).
The following line of summary output (
< 80$\text{}<80$ characters) is produced at every iteration. In all cases, the values of the quantities printed are those in effect
on completion of the given iteration.
Itn 
is the iteration count.

Step 
is the step taken along the computed search direction. If a constraint is added during the current iteration (i.e., Jadd is positive), Step will be the step to the nearest constraint. During the optimality phase, the step can be greater than one only if the factor R_{Z}${R}_{Z}$ is singular.
(See Section [Main Iteration].)

Ninf 
is the number of violated constraints (infeasibilities). This will be zero during the optimality phase.

Sinf/Objective 
is the value of the current objective function. If x$x$ is not feasible, Sinf gives a weighted sum of the magnitudes of constraint violations. If x$x$ is feasible, Objective is the value of the objective function of (1). The output line for the final iteration of the feasibility phase (i.e., the first iteration for which Ninf is zero) will give the value of the true objective at the first feasible point. During the optimality phase the value of the objective function will be nonincreasing. During the feasibility phase the number of constraint infeasibilities will not increase until either a feasible point is found or the optimality of the multipliers implies that no feasible point exists. Once optimal multipliers are obtained the number of infeasibilities can increase, but the sum of infeasibilities will either remain constant or be reduced until the minimum sum of infeasibilities is found.

Norm Gz 
is
‖Z_{1}^{T}g_{FR}‖
$\Vert {Z}_{1}^{\mathrm{T}}{g}_{\mathrm{FR}}\Vert $, the Euclidean norm of the reduced gradient with respect to Z_{1}${Z}_{1}$. During the optimality phase, this norm will be approximately zero after a unit step.
(See Sections [Definition of Search Direction] and [Main Iteration].)

The final printout includes a listing of the status of every variable and constraint.
The following describes the printout for each variable. A full stop (.) is printed for any numerical value that is zero.
Varbl 
gives the name (V) and index j$\mathit{j}$, for j = 1,2, … ,n$\mathit{j}=1,2,\dots ,n$, of the variable.

State 
gives the state of the variable (FR if neither bound is in the working set, EQ if a fixed variable, LL if on its lower bound, UL if on its upper bound, TF if temporarily fixed at its current value). If Value lies outside the upper or lower bounds by more than the Feasibility Tolerance, State will be ++ or  respectively.
A key is sometimes printed before State.
A 
Alternative optimum possible. The variable is active at one of its bounds, but its Lagrange multiplier is essentially zero. This means that if the variable were allowed to start moving away from its bound then there would be no change to the objective function. The values of the other free variables might change, giving a genuine alternative solution. However, if there are any degenerate variables (labelled D), the actual change might prove to be zero, since one of them could encounter a bound immediately. In either case the values of the Lagrange multipliers might also change.

D 
Degenerate. The variable is free, but it is equal to (or very close to) one of its bounds.

I 
Infeasible. The variable is currently violating one of its bounds by more than the Feasibility Tolerance.


Value 
is the value of the variable at the final iteration.

Lower Bound 
is the lower bound specified for the variable. None indicates that bl(j) ≤ − bigbnd${\mathbf{bl}}\left(j\right)\le \mathit{bigbnd}$.

Upper Bound 
is the upper bound specified for the variable. None indicates that bu(j) ≥ bigbnd${\mathbf{bu}}\left(j\right)\ge \mathit{bigbnd}$.

Lagr Mult 
is the Lagrange multiplier for the associated bound. This will be zero if State is FR unless bl(j) ≤ − bigbnd${\mathbf{bl}}\left(j\right)\le \mathit{bigbnd}$ and bu(j) ≥ bigbnd${\mathbf{bu}}\left(j\right)\ge \mathit{bigbnd}$, in which case the entry will be blank. If x$x$ is optimal, the multiplier should be nonnegative if State is LL and nonpositive if State is UL.

Slack 
is the difference between the variable Value and the nearer of its (finite) bounds bl(j)${\mathbf{bl}}\left(j\right)$ and bu(j)${\mathbf{bu}}\left(j\right)$. A blank entry indicates that the associated variable is not bounded (i.e., bl(j) ≤ − bigbnd${\mathbf{bl}}\left(j\right)\le \mathit{bigbnd}$ and bu(j) ≥ bigbnd${\mathbf{bu}}\left(j\right)\ge \mathit{bigbnd}$).

The meaning of the printout for general constraints is the same as that given above for variables, with ‘variable’ replaced by ‘constraint’,
bl(j)${\mathbf{bl}}\left(j\right)$ and
bu(j)${\mathbf{bu}}\left(j\right)$ are replaced by
bl(n + j)${\mathbf{bl}}\left(n+j\right)$ and
bu(n + j)${\mathbf{bu}}\left(n+j\right)$ respectively, and with the following change in the heading:
L Con 
gives the name (L) and index j$\mathit{j}$, for j = 1,2, … ,n_{L}$\mathit{j}=1,2,\dots ,{n}_{L}$, of the linear constraint.

Note that movement off a constraint (as opposed to a variable moving away from its bound) can be interpreted as allowing the entry in the Slack column to become positive.
Numerical values are output with a fixed number of digits; they are not guaranteed to be accurate to this precision.
Example
Open in the MATLAB editor:
nag_opt_lsq_lincon_solve_example
function nag_opt_lsq_lincon_solve_example
c = [1, 1, 1, 1, 1, 1, 1, 1, 4;
1, 2, 3, 4, 2, 1, 1, 1, 1;
1, 1, 1, 1, 1, 1, 1, 1, 1];
bl = [0;
0;
1e25;
0;
0;
0;
0;
0;
0;
2;
1e25;
1];
bu = [2;
2;
2;
2;
2;
2;
2;
2;
2;
1e25;
2;
4];
cvec = [0];
istate= zeros(12, 1, 'int64');
kx = zeros(9, 1, 'int64');
x = [1;
0.5;
0.3333;
0.25;
0.2;
0.1667;
0.1428;
0.125;
0.1111];
a = [1, 1, 1, 1, 1, 1, 1, 1, 1;
1, 2, 1, 1, 1, 1, 2, 0, 0;
1, 1, 3, 1, 1, 1, 1, 1, 3;
1, 1, 1, 4, 1, 1, 1, 1, 1;
1, 1, 1, 3, 1, 1, 1, 1, 1;
1, 1, 2, 1, 1, 0, 0, 0, 1;
1, 1, 1, 1, 0, 1, 1, 1, 1;
1, 1, 1, 0, 1, 1, 1, 1, 1;
1, 1, 0, 1, 1, 1, 2, 2, 3;
1, 0, 1, 1, 1, 1, 0, 2, 2 ];
b = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1,];
[cwsav,lwsav,iwsav,rwsav,ifail] = nag_opt_init('nag_opt_lsq_lincon_solve');
[istateOut, kxOut, xOut, aOut, bOut, iter, obj, clamda, lwsavOut, iwsavOut, rwsavOut, ifail] = ...
nag_opt_lsq_lincon_solve(c, bl, bu, cvec, istate, kx, x, a, b, lwsav, iwsav, rwsav);
istateOut, kxOut, xOut, aOut, bOut, iter, obj, clamda, ifail
istateOut =
1
0
0
1
0
1
0
1
0
1
2
1
kxOut =
3
9
2
5
7
6
8
1
4
xOut =
0
0.0415
0.5872
0
0.0996
0
0.0491
0
0.3056
aOut =
4.8017 0.2202 3.1226 0.4718 1.7539 2.1469 0.5193 2.5809 3.5110
0.1502 3.0381 0.3365 3.4674 2.1886 1.0342 1.6942 0.8582 1.4536
0.1502 0.6322 1.0175 1.0208 1.2471 0.1862 1.0728 0.2052 0.7820
0.6009 0.1227 0.3962 3.0130 2.8247 1.5832 3.1160 1.6001 2.2064
0.4507 0.0458 0.2083 0.1533 0.0000 0.7523 0.0000 0.0000 1.5022
0.1502 0.2620 0.2116 0.0654 0.0538 0.4627 0.0000 0.0000 3.1342
0.1502 0.1081 0.1676 0.1745 0.9495 0.0505 0.0000 0.0000 0.0000
0 0.1851 0.3556 0.3385 0.1728 0.0329 0.1766 0.0000 0.0000
0.1502 0.4783 0.1236 0.4145 0.1743 0.0118 0.4615 0.2966 0.0000
0.1502 0.2932 0.2715 0.1445 0.1684 0.0106 0.1221 0.2072 0.4538
bOut =
Columns 1 through 9
0.0000 0 0.2218 0.3369 0.0000 0.0000 0.0000 0.0000 0.0000
Column 10
0.0000
iter =
12
obj =
0.0813
clamda =
0.1572
0
0
0.8782
0
0.1473
0
0.8603
0
0.3777
0.0579
0.1075
ifail =
0
Open in the MATLAB editor:
e04nc_example
function e04nc_example
c = [1, 1, 1, 1, 1, 1, 1, 1, 4;
1, 2, 3, 4, 2, 1, 1, 1, 1;
1, 1, 1, 1, 1, 1, 1, 1, 1];
bl = [0;
0;
1e25;
0;
0;
0;
0;
0;
0;
2;
1e25;
1];
bu = [2;
2;
2;
2;
2;
2;
2;
2;
2;
1e25;
2;
4];
cvec = [0];
istate= zeros(12, 1, 'int64');
kx = zeros(9, 1, 'int64');
x = [1;
0.5;
0.3333;
0.25;
0.2;
0.1667;
0.1428;
0.125;
0.1111];
a = [1, 1, 1, 1, 1, 1, 1, 1, 1;
1, 2, 1, 1, 1, 1, 2, 0, 0;
1, 1, 3, 1, 1, 1, 1, 1, 3;
1, 1, 1, 4, 1, 1, 1, 1, 1;
1, 1, 1, 3, 1, 1, 1, 1, 1;
1, 1, 2, 1, 1, 0, 0, 0, 1;
1, 1, 1, 1, 0, 1, 1, 1, 1;
1, 1, 1, 0, 1, 1, 1, 1, 1;
1, 1, 0, 1, 1, 1, 2, 2, 3;
1, 0, 1, 1, 1, 1, 0, 2, 2 ];
b = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1,];
[cwsav,lwsav,iwsav,rwsav,ifail] = e04wb('e04nc');
[istateOut, kxOut, xOut, aOut, bOut, iter, obj, clamda, lwsavOut, iwsavOut, rwsavOut, ifail] = ...
e04nc(c, bl, bu, cvec, istate, kx, x, a, b, lwsav, iwsav, rwsav);
istateOut, kxOut, xOut, aOut, bOut, iter, obj, clamda, ifail
istateOut =
1
0
0
1
0
1
0
1
0
1
2
1
kxOut =
3
9
2
5
7
6
8
1
4
xOut =
0
0.0415
0.5872
0
0.0996
0
0.0491
0
0.3056
aOut =
4.8017 0.2202 3.1226 0.4718 1.7539 2.1469 0.5193 2.5809 3.5110
0.1502 3.0381 0.3365 3.4674 2.1886 1.0342 1.6942 0.8582 1.4536
0.1502 0.6322 1.0175 1.0208 1.2471 0.1862 1.0728 0.2052 0.7820
0.6009 0.1227 0.3962 3.0130 2.8247 1.5832 3.1160 1.6001 2.2064
0.4507 0.0458 0.2083 0.1533 0.0000 0.7523 0.0000 0.0000 1.5022
0.1502 0.2620 0.2116 0.0654 0.0538 0.4627 0.0000 0.0000 3.1342
0.1502 0.1081 0.1676 0.1745 0.9495 0.0505 0.0000 0.0000 0.0000
0 0.1851 0.3556 0.3385 0.1728 0.0329 0.1766 0.0000 0.0000
0.1502 0.4783 0.1236 0.4145 0.1743 0.0118 0.4615 0.2966 0.0000
0.1502 0.2932 0.2715 0.1445 0.1684 0.0106 0.1221 0.2072 0.4538
bOut =
Columns 1 through 9
0.0000 0 0.2218 0.3369 0.0000 0.0000 0.0000 0.0000 0.0000
Column 10
0.0000
iter =
12
obj =
0.0813
clamda =
0.1572
0
0
0.8782
0
0.1473
0
0.8603
0
0.3777
0.0579
0.1075
ifail =
0
Note: the remainder of this document is intended for more advanced users. Section [Algorithmic Details] contains a detailed description of the algorithm which may be needed in order to understand Sections [Optional Parameters] and [Description of Monitoring Information]. Section [Optional Parameters] describes the optional parameters which may be set by calls to nag_opt_lsq_lincon_option_string (e04ne). Section [Description of Monitoring Information] describes the quantities which can be requested to monitor the course of the computation.
Algorithmic Details
This section contains a detailed description of the method used by nag_opt_lsq_lincon_solve (e04nc).
Overview
nag_opt_lsq_lincon_solve (e04nc) is essentially identical to the function LSSOL described in
Gill et al. (1986). It is based on a twophase (primal) quadratic programming method with features to exploit the convexity of the objective function due to
Gill et al. (1984). (In the fullrank case, the method is related to that of
Stoer (1971).)
nag_opt_lsq_lincon_solve (e04nc) has two phases: finding an initial feasible point by minimizing the sum of infeasibilities (the
feasibility phase), and minimizing the quadratic objective function within the feasible region (the
optimality phase). The twophase nature of the algorithm is reflected by changing the function being minimized from the sum of infeasibilities to the quadratic objective function. The feasibility phase does
not perform the standard simplex method (i.e., it does not necessarily find a vertex), except in the LP case when
n_{L} ≤ n${n}_{L}\le n$. Once any iterate is feasible, all subsequent iterates remain feasible.
nag_opt_lsq_lincon_solve (e04nc) has been designed to be efficient when used to solve a
sequence of related problems – for example, within a sequential quadratic programming method for nonlinearly constrained optimization (e.g.,
nag_opt_nlp1_rcomm (e04uf) or
nag_opt_nlp2_solve (e04wd)). In particular, you may specify an initial working set (the indices of the constraints believed to be satisfied exactly at the solution); see the discussion of the optional parameter
Warm Start.
In general, an iterative process is required to solve a quadratic program. (For simplicity, we shall always consider a typical iteration and avoid reference to the index of the iteration.) Each new iterate
x$\stackrel{}{x}$ is defined by
where the
step length
α$\alpha $ is a nonnegative scalar, and
p$p$ is called the
search direction.At each point
x$x$, a
working set of constraints is defined to be a linearly independent subset of the constraints that are satisfied ‘exactly’ (to within the tolerance defined by the optional parameter
Feasibility Tolerance). The working set is the current prediction of the constraints that hold with equality at a solution of
(1). The search direction is constructed so that the constraints in the working set remain
unaltered for any value of the step length. For a bound constraint in the working set, this property is achieved by setting the corresponding element of the search direction to zero. Thus, the associated variable is
fixed, and specification of the working set induces a partition of
x$x$ into
fixed and
free variables. During a given iteration, the fixed variables are effectively removed from the problem; since the relevant elements of the search direction are zero, the columns of
c${\mathbf{c}}$ corresponding to fixed variables may be ignored.
Let
n_{W}${n}_{\mathrm{W}}$ denote the number of general constraints in the working set and let
n_{FX}${n}_{\mathrm{FX}}$ denote the number of variables fixed at one of their bounds (
n_{W}${n}_{\mathrm{W}}$ and
n_{FX}${n}_{\mathrm{FX}}$ are the quantities
Lin and
Bnd in the monitoring file output from
nag_opt_lsq_lincon_solve (e04nc); see
Section [Description of Monitoring Information]). Similarly, let
n_{FR}(n_{FR} = n − n_{FX})${n}_{\mathrm{FR}}({n}_{\mathrm{FR}}=n{n}_{\mathrm{FX}})$ denote the number of free variables. At every iteration,
the variables are reordered so that the last
n_{FX}${n}_{\mathrm{FX}}$ variables are fixed, with all other relevant vectors and matrices ordered accordingly. The order of the variables is indicated by the contents of the array
kx on exit (see
Section [Parameters]).
Definition of Search Direction
Let
C_{FR}${C}_{\mathrm{FR}}$ denote the
n_{W}${n}_{\mathrm{W}}$ by
n_{FR}${n}_{\mathrm{FR}}$ submatrix of general constraints in the working set corresponding to the free variables, and let
p_{FR}${p}_{\mathrm{FR}}$ denote the search direction with respect to the free variables only. The general constraints in the working set will be unaltered by any move along
p$p$ if
In order to compute
p_{FR}${p}_{\mathrm{FR}}$, the
TQ$TQ$ factorization of
C_{FR}${C}_{\mathrm{FR}}$ is used:
where
T$T$ is a nonsingular
n_{W}${n}_{\mathrm{W}}$ by
n_{W}${n}_{\mathrm{W}}$ reversetriangular matrix (i.e.,
t_{ij} = 0${t}_{ij}=0$ if
i + j < n_{W}$i+j<{n}_{\mathrm{W}}$), and the nonsingular
n_{FR}${n}_{\mathrm{FR}}$ by
n_{FR}${n}_{\mathrm{FR}}$ matrix
Q_{FR}${Q}_{\mathrm{FR}}$ is the product of orthogonal transformations (see
Gill et al. (1984)). If the columns of
Q_{FR}${Q}_{\mathrm{FR}}$ are partitioned so that
where
Y$Y$ is
n_{FR}${n}_{\mathrm{FR}}$ by
n_{W}${n}_{\mathrm{W}}$, then the
n_{Z}(n_{Z} = n_{FR} − n_{W})${n}_{Z}({n}_{Z}={n}_{\mathrm{FR}}{n}_{\mathrm{W}})$ columns of
Z$Z$ form a basis for the null space of
C_{FR}${C}_{\mathrm{FR}}$. Let
n_{R}${n}_{R}$ be an integer such that
0 ≤ n_{R} ≤ n_{Z}$0\le {n}_{R}\le {n}_{Z}$, and let
Z_{1}${Z}_{1}$ denote a matrix whose
n_{R}${n}_{R}$ columns are a subset of the columns of
Z$Z$. (The integer
n_{R}${n}_{R}$ is the quantity
Zr in the monitoring file output from
nag_opt_lsq_lincon_solve (e04nc). In many cases,
Z_{1}${Z}_{1}$ will include
all the columns of
Z$Z$.) The direction
p_{FR}${p}_{\mathrm{FR}}$ will satisfy
(3) if
where
p_{Z}${p}_{Z}$ is any
n_{R}${n}_{R}$vector.
Main Iteration
Let
Q$Q$ denote the
n$n$ by
n$n$ matrix
where
I_{FX}${I}_{\mathrm{FX}}$ is the identity matrix of order
n_{FX}${n}_{\mathrm{FX}}$. Let
R$R$ denote an
n$n$ by
n$n$ upper triangular matrix (the
Cholesky factor) such that
where
H̃$\stackrel{~}{H}$ is the Hessian
H$H$ with rows and columns permuted so that the free variables are first.
Let the matrix of the first
n_{Z}${n}_{Z}$ rows and columns of
R$R$ be denoted by
R_{Z}${R}_{Z}$. The definition of
p_{Z}${p}_{Z}$ in
(6) depends on whether or not the matrix
R_{Z}${R}_{Z}$ is singular at
x$x$. In the nonsingular case,
p_{Z}${p}_{Z}$ satisfies the equations
where
g_{Z}${g}_{Z}$ denotes the vector
Z^{T}g_{FR}${Z}^{\mathrm{T}}{g}_{\mathrm{FR}}$ and
g$g$ denotes the objective gradient. (The norm of
g_{FR}${g}_{\mathrm{FR}}$ is the printed quantity
Norm Gf; see
Section [Description of Monitoring Information].) When
p_{Z}${p}_{Z}$ is defined by
(9),
x + p$x+p$ is the minimizer of the objective function subject to the constraints (bounds and general) in the working set treated as equalities. In general, a vector
f_{Z}${f}_{Z}$ is available such that
R_{Z}^{T}
f_{Z}
=
−
g_{Z}
${R}_{Z}^{\mathrm{T}}{f}_{Z}={g}_{Z}$, which allows
p_{Z}${p}_{Z}$ to be computed from a single backsubstitution
R_{Z}p_{Z} = f_{Z}${R}_{Z}{p}_{Z}={f}_{Z}$. For example, when solving problem LS1,
f_{Z}${f}_{Z}$ comprises the first
n_{Z}${n}_{Z}$ elements of the
transformed residual vector
which is recurred from one iteration to the next, where
P$P$ is an orthogonal matrix.
In the singular case,
p_{Z}${p}_{Z}$ is defined such that
This vector has the property that the objective function is linear along
p$p$ and may be reduced by any step of the form
x + αp$x+\alpha p$, where
α > 0$\alpha >0$.
The vector
Z^{T}g_{FR}${Z}^{\mathrm{T}}{g}_{\mathrm{FR}}$ is known as the
projected gradient at
x$x$. If the projected gradient is zero,
x$x$ is a constrained stationary point in the subspace defined by
Z$Z$. During the feasibility phase, the projected gradient will usually be zero only at a vertex (although it may be zero at nonvertices in the presence of constraint dependencies). During the optimality phase, a zero projected gradient implies that
x$x$ minimizes the quadratic objective when the constraints in the working set are treated as equalities. At a constrained stationary point, Lagrange multipliers
λ_{c}${\lambda}_{{\mathbf{c}}}$ and
λ_{b}${\lambda}_{{\mathbf{b}}}$ for the general and bound constraints are defined from the equations
Given a positive constant
δ$\delta $ of the order of the
machine precision, the Lagrange multiplier
λ_{j}${\lambda}_{j}$ corresponding to an inequality constraint in the working set is said to be
optimal if
λ_{j} ≤ δ${\lambda}_{j}\le \delta $ when the associated constraint is at its
upper bound, or if
λ_{j} ≥ − δ${\lambda}_{j}\ge \delta $ when the associated constraint is at its
lower bound. If a multiplier is nonoptimal, the objective function (either the true objective or the sum of infeasibilities) can be reduced by deleting the corresponding constraint (with index
Jdel; see
Section [Description of Monitoring Information]) from the working set.
If optimal multipliers occur during the feasibility phase and the sum of infeasibilities is nonzero, there is no feasible point, and nag_opt_lsq_lincon_solve (e04nc) will continue until the minimum value of the sum of infeasibilities has been found. At this point, the Lagrange multiplier λ_{j}${\lambda}_{j}$ corresponding to an inequality constraint in the working set will be such that − (1 + δ) ≤ λ_{j} ≤ δ$(1+\delta )\le {\lambda}_{j}\le \delta $ when the associated constraint is at its upper bound, and − δ ≤ λ_{j} ≤ (1 + δ)$\delta \le {\lambda}_{j}\le (1+\delta )$ when the associated constraint is at its lower bound. Lagrange multipliers for equality constraints will satisfy λ_{j} ≤ 1 + δ$\left{\lambda}_{j}\right\le 1+\delta $.
The choice of step length is based on remaining feasible with respect to the satisfied constraints. If
R_{Z}${R}_{Z}$ is nonsingular and
x + p$x+p$ is feasible,
α$\alpha $ will be taken as unity. In this case, the projected gradient at
x$\stackrel{}{x}$ will be zero, and Lagrange multipliers are computed. Otherwise,
α$\alpha $ is set to
α_{m}${\alpha}_{{\mathbf{m}}}$, the step to the ‘nearest’ constraint (with index
Jadd; see
Section [Description of Monitoring Information]), which is added to the working set at the next iteration.
If
A$A$ is not input as a triangular matrix, it is overwritten by a triangular matrix
R$R$ satisfying
(8) obtained using the Cholesky factorization in the QP case, or the
QR$QR$ factorization in the LS case. Column interchanges are used in both cases, and an estimate is made of the rank of the triangular factor. Thereafter, the dependent rows of
R$R$ are eliminated from the problem.
Each change in the working set leads to a simple change to
C_{FR}${C}_{\mathrm{FR}}$: if the status of a general constraint changes, a
row of
C_{FR}${C}_{\mathrm{FR}}$ is altered; if a bound constraint enters or leaves the working set, a
column of
C_{FR}${C}_{\mathrm{FR}}$ changes. Explicit representations are recurred of the matrices
T,Q_{FR}$T,{Q}_{\mathrm{FR}}$ and
R$R$; and of vectors
Q^{T}g${Q}^{\mathrm{T}}g$,
Q^{T}c${Q}^{\mathrm{T}}c$ and
f$f$, which are related by the formulae
and
Note that the triangular factor
R$R$ associated with the Hessian of the original problem is updated during both the optimality
and the feasibility phases.
The treatment of the singular case depends critically on the following feature of the matrix updating schemes used in
nag_opt_lsq_lincon_solve (e04nc): if a given factor
R_{Z}${R}_{Z}$ is nonsingular, it can become singular during subsequent iterations only when a constraint leaves the working set, in which case only its last diagonal element can become zero. This property implies that a vector satisfying
(11) may be found using the single backsubstitution
R_{Z}p_{Z} = e_{Z}${\stackrel{}{R}}_{Z}{p}_{Z}={e}_{Z}$, where
R_{Z}${\stackrel{}{R}}_{Z}$ is the matrix
R_{Z}${R}_{Z}$ with a unit last diagonal, and
e_{Z}${e}_{Z}$ is a vector of all zeros except in the last position. If
H$H$ is singular, the matrix
R$R$ (and hence
R_{Z}${R}_{Z}$) may be singular at the start of the optimality phase. However,
R_{Z}${R}_{Z}$ will be nonsingular if enough constraints are included in the initial working set. (The matrix with no rows and columns is positive definite by definition, corresponding to the case when
C_{FR}${C}_{\mathrm{FR}}$ contains
n_{FR}${n}_{\mathrm{FR}}$ constraints.) The idea is to include as many general constraints as necessary to ensure a nonsingular
R_{Z}${R}_{Z}$.
At the beginning of each phase, an upper triangular matrix
R_{1}${R}_{1}$ is determined that is the largest nonsingular leading submatrix of
R_{Z}${R}_{Z}$. The use of interchanges during the factorization of
A$A$ tends to maximize the dimension of
R_{1}${R}_{1}$. (The rank of
R_{1}${R}_{1}$ is estimated using the optional parameter
Rank Tolerance.) Let
Z_{1}${Z}_{1}$ denote the columns of
Z$Z$ corresponding to
R_{1}${R}_{1}$, and let
Z$Z$ be partitioned as
Z = (Z_{1} Z_{2})$Z=\left({Z}_{1}\text{\hspace{1em}}{Z}_{2}\right)$. A working set for which
Z_{1}${Z}_{1}$ defines the null space can be obtained by including
the rows of
Z_{2}^{T}
${Z}_{2}^{\mathrm{T}}$ as ‘artificial constraints’. Minimization of the objective function then proceeds within the subspace defined by
Z_{1}${Z}_{1}$.
The artificially augmented working set is given by
so that
p_{FR}${p}_{\mathrm{FR}}$ will satisfy
C_{FR}p_{FR} = 0${C}_{\mathrm{FR}}{p}_{\mathrm{FR}}=0$ and
Z_{2}^{T}
p_{FR}
=
0
${Z}_{2}^{\mathrm{T}}{p}_{\mathrm{FR}}=0$. By definition of the
TQ$TQ$ factorization,
c_{FR}${\stackrel{}{{\mathbf{c}}}}_{\mathrm{FR}}$ automatically satisfies the following:
where
and hence the
TQ$TQ$ factorization of
(13) requires no additional work.
The matrix Z_{2}${Z}_{2}$ need not be kept fixed, since its role is purely to define an appropriate null space; the TQ$TQ$ factorization can therefore be updated in the normal fashion as the iterations proceed. No work is required to ‘delete’ the artificial constraints associated with Z_{2}${Z}_{2}$ when
Z_{1}^{T}
g_{FR}
=
0
${Z}_{1}^{\mathrm{T}}{g}_{\mathrm{FR}}=0$, since this simply involves repartitioning Q_{FR}${Q}_{\mathrm{FR}}$. When deciding which constraint to delete, the ‘artificial’ multiplier vector associated with the rows of
Z_{2}^{T}
${Z}_{2}^{\mathrm{T}}$ is equal to
Z_{2}^{T}
g_{FR}
${Z}_{2}^{\mathrm{T}}{g}_{\mathrm{FR}}$, and the multipliers corresponding to the rows of the ‘true’ working set are the multipliers that would be obtained if the temporary constraints were not present.
The number of columns in
Z_{2}${Z}_{2}$ and
Z_{1}
${Z}_{1}$, the Euclidean norm of
Z_{1}^{T}
g_{FR}
${Z}_{1}^{\mathrm{T}}{g}_{\mathrm{FR}}$, and the condition estimator of
R_{1}${R}_{1}$ appear in the monitoring file output as
Art,
Zr,
Norm Gz and
Cond Rz respectively (see
Section [Description of Monitoring Information]).
Although the algorithm of nag_opt_lsq_lincon_solve (e04nc) does not perform simplex steps in general, there is one exception: a linear program with fewer general constraints than variables (i.e., n_{L} ≤ n${n}_{L}\le n$). Use of the simplex method in this situation leads to savings in storage. At the starting point, the ‘natural’ working set (the set of constraints exactly or nearly satisfied at the starting point) is augmented with a suitable number of ‘temporary’ bounds, each of which has the effect of temporarily fixing a variable at its current value. In subsequent iterations, a temporary bound is treated as a standard constraint until it is deleted from the working set, in which case it is never added again.
One of the most important features of
nag_opt_lsq_lincon_solve (e04nc) is its control of the conditioning of the working set, whose nearness to linear dependence is estimated by the ratio of the largest to smallest diagonals of the
TQ$TQ$ factor
T$T$ (the printed value
Cond T; see
Section [Description of Monitoring Information]). In constructing the initial working set, constraints are excluded that would result in a large value of
Cond T. Thereafter,
nag_opt_lsq_lincon_solve (e04nc) allows constraints to be violated by as much as a userspecified optional parameter
Feasibility Tolerance in order to provide, whenever possible, a
choice of constraints to be added to the working set at a given iteration. Let
α_{m}${\alpha}_{{\mathbf{m}}}$ denote the maximum step at which
x + α_{m}p$x+{\alpha}_{{\mathbf{m}}}p$ does not violate any constraint by more than its feasibility tolerance. All constraints at distance
α(α ≤ α_{m})$\alpha (\alpha \le {\alpha}_{{\mathbf{m}}})$ along
p$p$ from the current point are then viewed as acceptable candidates for inclusion in the working set. The constraint whose normal makes the largest angle with the search direction is added to the working set. In order to ensure that the new iterate satisfies the constraints in the working set as accurately as possible, the step taken is the exact distance to the newly added constraint. As a consequence, negative steps are occasionally permitted, since the current iterate may violate the constraint to be added by as much as the feasibility tolerance.
Optional Parameters
Several optional parameters in nag_opt_lsq_lincon_solve (e04nc) define choices in the problem specification or the algorithm logic. In order to reduce the number of formal parameters of nag_opt_lsq_lincon_solve (e04nc) these optional parameters have associated default values that are appropriate for most problems. Therefore, you need only specify those optional parameters whose values are to be different from their default values.
The remainder of this section can be skipped if you wish to use the default values for all optional parameters.
The following is a list of the optional parameters available. A full description of each optional parameter is provided in
Section [Description of the optional parameters].
Optional parameters may be specified by calling
nag_opt_lsq_lincon_option_string (e04ne) before a call to
nag_opt_lsq_lincon_solve (e04nc).
nag_opt_lsq_lincon_option_string (e04ne) can be called to supply options directly, one call being necessary for each optional parameter. For example,
[lwsav, iwsav, rwsav, inform] = e04ne('Print Level = 1', lwsav, iwsav, rwsav);
nag_opt_lsq_lincon_option_string (e04ne) should be consulted for a full description of this method of supplying optional parameters.
All optional parameters not specified by you are set to their default values. Optional parameters specified by you are unaltered by nag_opt_lsq_lincon_solve (e04nc) (unless they define invalid values) and so remain in effect for subsequent calls unless altered by you.
Description of the Optional Parameters
For each option, we give a summary line, a description of the optional parameter and details of constraints.
The summary line contains:
 the keywords, where the minimum abbreviation of each keyword is underlined (if no characters of an optional qualifier are underlined, the qualifier may be omitted);
 a parameter value,
where the letters a$a$, i and r$i\text{ and}r$ denote options that take character, integer and real values respectively;
 the default value, where the symbol ε$\epsilon $ is a generic notation for machine precision (see nag_machine_precision (x02aj)).
Keywords and character values are case and white space insensitive.
Cold Start DefaultWarm Start This option specifies how the initial working set is chosen. With a
Cold Start,
nag_opt_lsq_lincon_solve (e04nc) chooses the initial working set based on the values of the variables and constraints at the initial point. Broadly speaking, the initial working set will include equality constraints and bounds or inequality constraints that violate or ‘nearly’ satisfy their bounds (to within
Crash Tolerance).
With a
Warm Start, you must provide a valid definition of every element of the array
istate.
nag_opt_lsq_lincon_solve (e04nc) will override your specification of
istate if necessary, so that a poor choice of the working set will not cause a fatal error. For instance, any elements of
istate which are set to
− 2$2$,
− 1 or 4$1\text{ or}4$ will be reset to zero, as will any elements which are set to
3$3$ when the corresponding elements of
bl and
bu are not equal. A warm start will be advantageous if a good estimate of the initial working set is available – for example, when
nag_opt_lsq_lincon_solve (e04nc) is called repeatedly to solve related problems.
Crash Tolerance r$r$Default = 0.01$\text{}=0.01$This value is used in conjunction with the optional parameter
Cold Start (the default value) when
nag_opt_lsq_lincon_solve (e04nc) selects an initial working set. If
0 ≤ r ≤ 1$0\le r\le 1$, the initial working set will include (if possible) bounds or general inequality constraints that lie within
r$r$ of their bounds. In particular, a constraint of the form
c_{j}^{T}
x ≥ l
${c}_{j}^{\mathrm{T}}x\ge l$ will be included in the initial working set if
c_{j}^{T}x − l
≤
r
(1 + l)
${c}_{j}^{\mathrm{T}}xl\le r(1+\leftl\right)$. If
r < 0$r<0$ or
r > 1$r>1$, the default value is used.
Defaults This special keyword may be used to reset all optional parameters to their default values.
Feasibility Phase Iteration Limit i_{1}${i}_{1}$Default = max (50,5(n + n_{L}))$\text{}=\mathrm{max}\phantom{\rule{0.125em}{0ex}}(50,5(n+{n}_{L}))$Optimality Phase Iteration Limit i_{2}${i}_{2}$Default = max (50,5(n + n_{L}))$\text{}=\mathrm{max}\phantom{\rule{0.125em}{0ex}}(50,5(n+{n}_{L}))$The scalars
i_{1}${i}_{1}$ and
i_{2}${i}_{2}$ specify the maximum number of iterations allowed in the feasibility and optimality phases. Optional parameter
Optimality Phase Iteration Limit is equivalent to optional parameter
Iteration Limit. Setting
i_{2} = 0${i}_{2}=0$ and
Print Level > 0${\mathbf{Print\; Level}}>0$ means that the workspace needed will be computed and printed, but no iterations will be performed. If
i_{1} < 0${i}_{1}<0$ or
i_{2} < 0${i}_{2}<0$, the default value is used.
Feasibility Tolerance r$r$Default = sqrt(ε)$\text{}=\sqrt{\epsilon}$If r > ε$r>\epsilon $, r$r$ defines the maximum acceptable absolute violation in each constraint at a ‘feasible’ point. For example, if the variables and the coefficients in the general constaints are of order unity, and the latter are correct to about 6$6$ decimal digits, it would be appropriate to specify r$r$ as 10^{ − 6}${10}^{6}$. If 0 ≤ r < ε$0\le r<\epsilon $, the default value is used.
Note that a ‘feasible solution’ is a solution that satisfies the current constraints to within the tolerance r$r$.
Hessian No$\overline{)\mathbf{N}}\mathbf{o}$Default = NO$\text{}=\mathrm{NO}$ This option controls the contents of the upper triangular matrix
R$R$ (see the description of
a in
Section [Parameters]).
nag_opt_lsq_lincon_solve (e04nc) works exclusively with the transformed and reordered matrix
H_{Q}${H}_{Q}$ (8), and hence extra computation is required to form the Hessian itself. If
Hessian = NO${\mathbf{Hessian}}=\mathrm{NO}$,
a contains the Cholesky factor of the matrix
H_{Q}${H}_{Q}$ with columns ordered as indicated by
kx (see
Section [Parameters]). If
Hessian = YES${\mathbf{Hessian}}=\mathrm{YES}$,
a contains the Cholesky factor of the matrix
H$H$, with columns ordered as indicated by
kx.
Infinite Bound Size r$r$Default = 10^{20}$\text{}={10}^{20}$If r > 0$r>0$, r$r$ defines the ‘infinite’ bound bigbnd$\mathit{bigbnd}$ in the definition of the problem constraints. Any upper bound greater than or equal to bigbnd$\mathit{bigbnd}$ will be regarded as + ∞$+\infty $ (and similarly any lower bound less than or equal to − bigbnd$\mathit{bigbnd}$ will be regarded as − ∞$\infty $). If r < 0$r<0$, the default value is used.
Infinite Step Size r$r$Default = max (bigbnd,10^{20})$\text{}=\mathrm{max}\phantom{\rule{0.125em}{0ex}}(\mathit{bigbnd},{10}^{20})$If r > 0$r>0$, r$r$ specifies the magnitude of the change in variables that will be considered a step to an unbounded solution. (Note that an unbounded solution can occur only when the Hessian is singular and the objective contains an explicit linear term.) If the change in x$x$ during an iteration would exceed the value of r$r$, the objective function is considered to be unbounded below in the feasible region. If r ≤ 0$r\le 0$, the default value is used.
Iteration Limit i$i$Default = max (50,5(n + n_{L}))$\text{}=\mathrm{max}\phantom{\rule{0.125em}{0ex}}(50,5(n+{n}_{L}))$Iters Itns List Default for e04nc = List$\text{e04nc}={\mathbf{List}}$Nolist Default for e04nc = Nolist$\text{e04nc}={\mathbf{Nolist}}$Normally each optional parameter specification is printed as it is supplied. Optional parameter
Nolist may be used to suppress the printing and optional parameter
List may be used to restore printing.
Monitoring File i$i$Default = − 1$\text{}=1$If
i ≥ 0$i\ge 0$ and
Print Level ≥ 5${\mathbf{Print\; Level}}\ge 5$, monitoring information produced by
nag_opt_lsq_lincon_solve (e04nc) at every iteration is sent to a file with logical unit number
i$i$. If
i < 0$i<0$ and/or
Print Level < 5${\mathbf{Print\; Level}}<5$, no monitoring information is produced.
Print Level i$i$The value of
i$i$ controls the amount of printout produced by
nag_opt_lsq_lincon_solve (e04nc), as indicated below. A detailed description of the printed output is given in
Section [Printed output] (summary output at each iteration and the final solution) and
Section [Description of Monitoring Information] (monitoring information at each iteration).
The following printout is sent to the current advisory message unit (as defined by
nag_file_set_unit_advisory (x04ab)):
i$i$ 
Output 
≥ 00$\phantom{\ge 0}0$ 
No output. 
≥ 01$\phantom{\ge 0}1$ 
The final solution only. 
≥ 05$\phantom{\ge 0}5$ 
One line of summary output ( < 80$\text{}<80$ characters; see Section [Printed output]) for each iteration (no printout of the final solution). 
≥ 10$\text{}\ge 10$ 
The final solution and one line of summary output for each iteration. 
The following printout is sent to the logical unit number defined by the optional parameter
Monitoring File:
i$i$ 
Output 
< 5$\text{}<5$ 
No output. 
≥ 5$\text{}\ge 5$ 
One long line of output ( > 80$\text{}>80$ characters; see Section [Description of Monitoring Information]) for each iteration (no printout of the final solution). 
≥ 20$\text{}\ge 20$ 
At each iteration, the Lagrange multipliers, the variables x$x$, the constraint values Cx$Cx$ and the constraint status. 
≥ 30$\text{}\ge 30$ 
At each iteration, the diagonal elements of the matrix T$T$ associated with the TQ$TQ$ factorization (4) (see Section [Definition of Search Direction]) of the working set, and the diagonal elements of the upper triangular matrix R$R$. 
If
Print Level ≥ 5${\mathbf{Print\; Level}}\ge 5$ and the unit number defined by the optional parameter
Monitoring File is the same as that defined by
nag_file_set_unit_advisory (x04ab), then the summary output is suppressed.
Problem Type a$a$Default = $=$ LS1This option specifies the type of objective function to be minimized during the optimality phase. The following are the nine optional keywords and the dimensions of the arrays that must be specified in order to define the objective function:
LP 
a and b not referenced, cvec(n)${\mathbf{cvec}}\left({\mathbf{n}}\right)$; 
QP1 
a(lda,n)${\mathbf{a}}\left(\mathit{lda},{\mathbf{n}}\right)$ symmetric, b and cvec not referenced; 
QP2 
a(lda,n)${\mathbf{a}}\left(\mathit{lda},{\mathbf{n}}\right)$ symmetric, b not referenced, cvec(n)${\mathbf{cvec}}\left({\mathbf{n}}\right)$; 
QP3 
a(lda,n)${\mathbf{a}}\left(\mathit{lda},{\mathbf{n}}\right)$ upper trapezoidal, kx(n)${\mathbf{kx}}\left({\mathbf{n}}\right)$, b and cvec not referenced; 
QP4 
a(lda,n)${\mathbf{a}}\left(\mathit{lda},{\mathbf{n}}\right)$ upper trapezoidal, kx(n)${\mathbf{kx}}\left({\mathbf{n}}\right)$, b not referenced, cvec(n)${\mathbf{cvec}}\left({\mathbf{n}}\right)$; 
LS1 
a(lda,n)${\mathbf{a}}\left(\mathit{lda},{\mathbf{n}}\right)$, b(m)${\mathbf{b}}\left({\mathbf{m}}\right)$, cvec not referenced; 
LS2 
a(lda,n)${\mathbf{a}}\left(\mathit{lda},{\mathbf{n}}\right)$, b(m)${\mathbf{b}}\left({\mathbf{m}}\right)$, cvec(n)${\mathbf{cvec}}\left({\mathbf{n}}\right)$; 
LS3 
a(lda,n)${\mathbf{a}}\left(\mathit{lda},{\mathbf{n}}\right)$ upper trapezoidal, kx(n)${\mathbf{kx}}\left({\mathbf{n}}\right)$, b(m)${\mathbf{b}}\left({\mathbf{m}}\right)$, cvec not referenced; 
LS4 
a(lda,n)${\mathbf{a}}\left(\mathit{lda},{\mathbf{n}}\right)$ upper trapezoidal, kx(n)${\mathbf{kx}}\left({\mathbf{n}}\right)$, b(m)${\mathbf{b}}\left({\mathbf{m}}\right)$, cvec(n)${\mathbf{cvec}}\left({\mathbf{n}}\right)$. 
For problems of type FP, the objective function is omitted and
a,
b and
cvec are not referenced.
The following keywords are also acceptable. The minimum abbreviation of each keyword is underlined.
a$a$ 
Option 
Least 
LS1 
Quadratic 
QP2 
Linear 
LP 
In addition, the keywords LS and LSQ are equivalent to the default option LS1, and the keyword QP is equivalent to the option QP2.
If A = 0$A=0$, i.e., the objective function is purely linear, the efficiency of nag_opt_lsq_lincon_solve (e04nc) may be increased by specifying a$a$ as LP.
Rank Tolerance r$r$Default = 100ε$\text{}=100\epsilon $ or 10sqrt(ε)$10\sqrt{\epsilon}$ (see below)Note that this option does not apply to problems of type FP or LP.
The default value of r$r$ depends on the problem type. If A$A$ occurs as a least squares matrix, as it does in problem types QP1, LS1 and LS3, then the default value of r$r$ is 100ε$100\epsilon $. In all other cases, A$A$ is treated as the ‘square root’ of the Hessian matrix H$H$ and r$r$ has the default value 10sqrt(ε)$10\sqrt{\epsilon}$.
This parameter enables you to control the estimate of the triangular factor
R_{1}${R}_{1}$ (see
Section [Main Iteration]). If
ρ_{i}${\rho}_{i}$ denotes the function
ρ_{i} = max {R_{11},R_{22}, … ,R_{ii}}${\rho}_{i}=\mathrm{max}\phantom{\rule{0.125em}{0ex}}\{\left{R}_{11}\right,\left{R}_{22}\right,\dots ,\left{R}_{ii}\right\}$, the rank of
R$R$ is defined to be smallest index
i such that
R_{i + 1,i + 1} ≤ rρ_{i + 1}$\left{R}_{i+1,i+1}\right\le r\left{\rho}_{i+1}\right$. If
r ≤ 0$r\le 0$, the default value is used.
Description of Monitoring Information
This section describes the long line of output (
> 80$\text{}>80$ characters) which forms part of the monitoring information produced by
nag_opt_lsq_lincon_solve (e04nc). (See also the description of the optional parameters
Monitoring File and
Print Level.)
You can control the level of printed output.
To aid interpretation of the printed results, the following convention is used for numbering the constraints: indices 1$1$ through n$n$ refer to the bounds on the variables, and indices n + 1$n+1$ through n + n_{L}$n+{n}_{L}$ refer to the general constraints. When the status of a constraint changes, the index of the constraint is printed, along with the designation L (lower bound), U (upper bound), E (equality), F (temporarily fixed variable) or A (artificial constraint).
When
Print Level ≥ 5${\mathbf{Print\; Level}}\ge 5$ and
Monitoring File ≥ 0${\mathbf{Monitoring\; File}}\ge 0$, the following line of output is produced at every iteration on the unit number specified by optional parameter
Monitoring File. In all cases, the values of the quantities printed are those in effect
on completion of the given iteration.
Itn 
is the iteration count.

Jdel 
is the index of the constraint deleted from the working set. If Jdel is zero, no constraint was deleted.

Jadd 
is the index of the constraint added to the working set. If Jadd is zero, no constraint was added.

Step 
is the step taken along the computed search direction. If a constraint is added during the current iteration (i.e., Jadd is positive), Step will be the step to the nearest constraint. During the optimality phase, the step can be greater than one only if the factor R_{Z}${R}_{Z}$ is singular.

Ninf 
is the number of violated constraints (infeasibilities). This will be zero during the optimality phase.

Sinf/Objective 
is the value of the current objective function. If x$x$ is not feasible, Sinf gives a weighted sum of the magnitudes of constraint violations. If x$x$ is feasible, Objective is the value of the objective function of (1). The output line for the final iteration of the feasibility phase (i.e., the first iteration for which Ninf is zero) will give the value of the true objective at the first feasible point. During the optimality phase the value of the objective function will be nonincreasing. During the feasibility phase the number of constraint infeasibilities will not increase until either a feasible point is found or the optimality of the multipliers implies that no feasible point exists. Once optimal multipliers are obtained the number of infeasibilities can increase, but the sum of infeasibilities will either remain constant or be reduced until the minimum sum of infeasibilities is found.

Bnd 
is the number of simple bound constraints in the current working set.

Lin 
is the number of general linear constraints in the current working set.

Art 
is the number of artificial constraints in the working set, i.e., the number of columns of Z_{2}${Z}_{2}$ (see Section [Main Iteration]).

Zr 
is the number of columns of Z_{1}${Z}_{1}$(see Section [Definition of Search Direction]). Zr is the dimension of the subspace in which the objective function is currently being minimized. The value of Zr is the number of variables minus the number of constraints in the working set; i.e., Zr = n − (Bnd + Lin + Art)$\mathtt{Zr}=n(\mathtt{Bnd}+\mathtt{Lin}+\mathtt{Art})$.The value of n_{Z}${n}_{Z}$, the number of columns of Z$Z$ (see Section [Definition of Search Direction]) can be calculated as n_{Z} = n − (Bnd + Lin)${n}_{Z}=n(\mathtt{Bnd}+\mathtt{Lin})$. A zero value of n_{Z}${n}_{Z}$ implies that x$x$ lies at a vertex of the feasible region.

Norm Gz 
is
‖Z_{1}^{T}g_{FR}‖
$\Vert {Z}_{1}^{\mathrm{T}}{g}_{\mathrm{FR}}\Vert $, the Euclidean norm of the reduced gradient with respect to Z_{1}${Z}_{1}$. During the optimality phase, this norm will be approximately zero after a unit step.

Norm Gf 
is the Euclidean norm of the gradient function with respect to the free variables, i.e., variables not currently held at a bound.

Cond T 
is a lower bound on the condition number of the working set.

Cond Rz 
is a lower bound on the condition number of the triangular factor R_{1}${R}_{1}$ (the first Zr rows and columns of the factor R_{Z}${R}_{Z}$). If the problem is specified to be of type LP or the estimated rank of the data matrix A$A$ is zero then Cond Rz is not printed.

PDF version (NAG web site
, 64bit version, 64bit version)
© The Numerical Algorithms Group Ltd, Oxford, UK. 2009–2013