NAG Library Routine Document
E02GBF
1 Purpose
E02GBF calculates an ${l}_{1}$ solution to an overdetermined system of linear equations, possibly subject to linear inequality constraints.
2 Specification
SUBROUTINE E02GBF ( 
M, N, MPL, E, LDE, F, X, MXS, MONIT, IPRINT, K, EL1N, INDX, W, IW, IFAIL) 
INTEGER 
M, N, MPL, LDE, MXS, IPRINT, K, INDX(MPL), IW, IFAIL 
REAL (KIND=nag_wp) 
E(LDE,MPL), F(MPL), X(N), EL1N, W(IW) 
EXTERNAL 
MONIT 

3 Description
Given a matrix
$A$ with
$m$ rows and
$n$ columns
$\left(m\ge n\right)$ and a vector
$b$ with
$m$ elements, the routine calculates an
${l}_{1}$ solution to the overdetermined system of equations
That is to say, it calculates a vector
$x$, with
$n$ elements, which minimizes the
${l}_{1}$norm (the sum of the absolute values) of the residuals
where the residuals
${r}_{i}$ are given by
Here
${a}_{ij}$ is the element in row
$i$ and column
$j$ of
$A$,
${b}_{i}$ is the
$i$th element of
$b$ and
${x}_{j}$ the
$j$th element of
$x$.
If, in addition, a matrix $C$ with $l$ rows and $n$ columns and a vector $d$ with $l$ elements, are given, the vector $x$ computed by the routine is such as to minimize the ${l}_{1}$norm $r\left(x\right)$ subject to the set of inequality constraints $Cx\ge d$.
The matrices $A$ and $C$ need not be of full rank.
Typically in applications to data fitting, data consisting of
$m$ points with coordinates
$\left({t}_{i},{y}_{i}\right)$ is to be approximated by a linear combination of known functions
${\varphi}_{i}\left(t\right)$,
in the
${l}_{1}$norm, possibly subject to linear inequality constraints on the coefficients
${\alpha}_{j}$ of the form
$C\alpha \ge d$ where
$\alpha $ is the vector of the
${\alpha}_{j}$ and
$C$ and
$d$ are as in the previous paragraph. This is equivalent to finding an
${l}_{1}$ solution to the overdetermined system of equations
subject to
$C\alpha \ge d$.
Thus if, for each value of $i$ and $j$, the element ${a}_{ij}$ of the matrix $A$ above is set equal to the value of ${\varphi}_{j}\left({t}_{i}\right)$ and ${b}_{i}$ is equal to ${y}_{i}$ and $C$ and $d$ are also supplied to the routine, the solution vector $x$ will contain the required values of the ${\alpha}_{j}$. Note that the independent variable $t$ above can, instead, be a vector of several independent variables (this includes the case where each of ${\varphi}_{i}$ is a function of a different variable, or set of variables).
The algorithm follows the Conn–Pietrzykowski approach (see
Bartels et al. (1978) and
Conn and Pietrzykowski (1977)), which is via an exact penalty function
where
$\gamma $ is a penalty parameter,
${c}_{i}^{\mathrm{T}}$ is the
$i$th row of the matrix
$C$, and
${d}_{i}$ is the
$i$th element of the vector
$d$. It proceeds in a stepbystep manner much like the simplex method for linear programming but does not move from vertex to vertex and does not require the problem to be cast in a form containing only nonnegative unknowns. It uses stable procedures to update an orthogonal factorization of the current set of active equations and constraints.
4 References
Bartels R H, Conn A R and Charalambous C (1976) Minimisation techniques for piecewise Differentiable functions – the ${l}_{\infty}$ solution to an overdetermined linear system Technical Report No. 247, CORR 76/30 Mathematical Sciences Department, The John Hopkins University
Bartels R H, Conn A R and Sinclair J W (1976) A Fortran program for solving overdetermined systems of linear equations in the ${l}_{1}$ Sense Technical Report No. 236, CORR 76/7 Mathematical Sciences Department, The John Hopkins University
Bartels R H, Conn A R and Sinclair J W (1978) Minimisation techniques for piecewise differentiable functions – the ${l}_{1}$ solution to an overdetermined linear system SIAM J. Numer. Anal. 15 224–241
Conn A R and Pietrzykowski T (1977) A penaltyfunction method converging directly to a constrained optimum SIAM J. Numer. Anal. 14 348–375
5 Parameters
 1: M – INTEGERInput
On entry: the number of equations in the overdetermined system, $m$ (i.e., the number of rows of the matrix $A$).
Constraint:
${\mathbf{M}}\ge 2$.
 2: N – INTEGERInput
On entry: the number of unknowns, $n$ (the number of columns of the matrix $A$).
Constraint:
${\mathbf{N}}\ge 2$.
 3: MPL – INTEGERInput
On entry: $m+l$, where $l$ is the number of constraints (which may be zero).
Constraint:
${\mathbf{MPL}}\ge {\mathbf{M}}$.
 4: E(LDE,MPL) – REAL (KIND=nag_wp) arrayInput/Output
On entry: the equation and constraint matrices stored in the following manner.
The first $m$ columns contain the $m$ rows of the matrix $A$; element
${\mathbf{E}}\left(\mathit{i},\mathit{j}\right)$ specifying the element ${a}_{\mathit{j}\mathit{i}}$ in the $\mathit{j}$th row and $\mathit{i}$th column of $A$ (the coefficient of the $\mathit{i}$th unknown in the $\mathit{j}$th equation), for $\mathit{i}=1,2,\dots ,n$ and $\mathit{j}=1,2,\dots ,m$. The next $l$ columns contain the $l$ rows of the constraint matrix $C$; element
${\mathbf{E}}\left(\mathit{i},\mathit{j}+m\right)$ containing the element ${c}_{\mathit{j}\mathit{i}}$ in the $\mathit{j}$th row and $\mathit{i}$th column of $C$ (the coefficient of the $\mathit{i}$th unknown in the $\mathit{j}$th constraint), for $\mathit{i}=1,2,\dots ,n$ and $\mathit{j}=1,2,\dots ,l$.
On exit: unchanged, except possibly to the extent of a small multiple of the
machine precision. (See
Section 8.)
 5: LDE – INTEGERInput
On entry: the first dimension of the array
E as declared in the (sub)program from which E02GBF is called.
Constraint:
${\mathbf{LDE}}\ge {\mathbf{N}}$.
 6: F(MPL) – REAL (KIND=nag_wp) arrayInput
On entry: ${\mathbf{F}}\left(\mathit{i}\right)$, for $\mathit{i}=1,2,\dots ,m$, must contain
${b}_{\mathit{i}}$ (the $\mathit{i}$th element of the righthand side vector of the overdetermined system of equations) and ${\mathbf{F}}\left(m+\mathit{i}\right)$, for $\mathit{i}=1,2,\dots ,l$, must contain ${d}_{i}$ (the $i$th element of the righthand side vector of the constraints), where $l$ is the number of constraints.
 7: X(N) – REAL (KIND=nag_wp) arrayInput/Output
On entry: ${\mathbf{X}}\left(\mathit{i}\right)$ must contain an estimate of the $\mathit{i}$th unknown, for $\mathit{i}=1,2,\dots ,n$. If no better initial estimate for ${\mathbf{X}}\left(i\right)$ is available, set ${\mathbf{X}}\left(i\right)=0.0$.
On exit: the latest estimate of the
$\mathit{i}$th unknown, for $\mathit{i}=1,2,\dots ,n$. If ${\mathbf{IFAIL}}={\mathbf{0}}$ on exit, these are the solution values.
 8: MXS – INTEGERInput
On entry: the maximum number of steps to be allowed for the solution of the unconstrained problem. Typically this may be a modest multiple of
$n$. If, on entry,
MXS is zero or negative, the value returned by
X02BBF is used.
 9: MONIT – SUBROUTINE, supplied by the user.External Procedure
MONIT can be used to print out the current values of any selection of its parameters. The frequency with which
MONIT is called in E02GBF is controlled by
IPRINT.
The specification of
MONIT is:
INTEGER 
N, NITER, K 
REAL (KIND=nag_wp) 
X(N), EL1N 

 1: N – INTEGERInput
On entry: the number $n$ of unknowns (the number of columns of the matrix $A$).
 2: X(N) – REAL (KIND=nag_wp) arrayInput
On entry: the latest estimate of the unknowns.
 3: NITER – INTEGERInput
On entry: the number of iterations so far carried out.
 4: K – INTEGERInput
On entry: the total number of equations and constraints which are currently active (i.e., the number of equations with zero residuals plus the number of constraints which are satisfied as equations).
 5: EL1N – REAL (KIND=nag_wp)Input
On entry: the ${l}_{1}$norm of the current residuals of the overdetermined system of equations.
MONIT must either be a module subprogram USEd by, or declared as EXTERNAL in, the (sub)program from which E02GBF is called. Parameters denoted as
Input must
not be changed by this procedure.
 10: IPRINT – INTEGERInput
On entry: the frequency of iteration print out.
 ${\mathbf{IPRINT}}>0$
 MONIT is called every IPRINT iterations and at the solution.
 ${\mathbf{IPRINT}}=0$
 Information is printed out at the solution only. Otherwise MONIT is not called (but a dummy routine must still be provided).
 11: K – INTEGEROutput
On exit: the total number of equations and constraints which are then active (i.e., the number of equations with zero residuals plus the number of constraints which are satisfied as equalities).
 12: EL1N – REAL (KIND=nag_wp)Output
On exit: the ${l}_{1}$norm (sum of absolute values) of the equation residuals.
 13: INDX(MPL) – INTEGER arrayOutput
On exit: specifies which columns of
E relate to the inactive equations and constraints.
${\mathbf{INDX}}\left(1\right)$ up to
${\mathbf{INDX}}\left({\mathbf{K}}\right)$ number the active columns and
${\mathbf{INDX}}\left({\mathbf{K}}+1\right)$ up to
${\mathbf{INDX}}\left({\mathbf{MPL}}\right)$ number the inactive columns.
 14: W(IW) – REAL (KIND=nag_wp) arrayWorkspace
 15: IW – INTEGERInput
On entry: the dimension of the array
W as declared in the (sub)program from which E02GBF is called.
Constraint:
${\mathbf{IW}}\ge 3\times {\mathbf{MPL}}+5\times {\mathbf{N}}+{{\mathbf{N}}}^{2}+\left({\mathbf{N}}+1\right)\times \left({\mathbf{N}}+2\right)/2$.
 16: IFAIL – INTEGERInput/Output

On entry:
IFAIL must be set to
$0$,
$1\text{ or}1$. If you are unfamiliar with this parameter you should refer to
Section 3.3 in the Essential Introduction for details.
For environments where it might be inappropriate to halt program execution when an error is detected, the value
$1\text{ or}1$ is recommended. If the output of error messages is undesirable, then the value
$1$ is recommended. Otherwise, if you are not familiar with this parameter, the recommended value is
$0$.
When the value $\mathbf{1}\text{ or}\mathbf{1}$ is used it is essential to test the value of IFAIL on exit.
On exit:
${\mathbf{IFAIL}}={\mathbf{0}}$ unless the routine detects an error or a warning has been flagged (see
Section 6).
6 Error Indicators and Warnings
If on entry
${\mathbf{IFAIL}}={\mathbf{0}}$ or
${{\mathbf{1}}}$, explanatory error messages are output on the current error message unit (as defined by
X04AAF).
Errors or warnings detected by the routine:
 ${\mathbf{IFAIL}}=1$
The constraints cannot all be satisfied simultaneously: they are not compatible with one another. Hence no solution is possible.
 ${\mathbf{IFAIL}}=2$
The limit imposed by
MXS has been reached without finding a solution. Consider restarting from the current point by simply calling E02GBF again without changing the parameters.
 ${\mathbf{IFAIL}}=3$
The routine has failed because of numerical difficulties; the problem is too illconditioned. Consider rescaling the unknowns.
 ${\mathbf{IFAIL}}=4$
On entry, one or more of the following conditions are violated:
 ${\mathbf{M}}\ge {\mathbf{N}}\ge 2$,
 or ${\mathbf{MPL}}\ge {\mathbf{M}}$,
 or ${\mathbf{IW}}\ge 3\times {\mathbf{MPL}}+5\times {\mathbf{N}}+{{\mathbf{N}}}^{2}+\left({\mathbf{N}}+1\right)\times \left({\mathbf{N}}+2\right)/2$,
 or ${\mathbf{LDE}}\ge {\mathbf{N}}$.
Alternatively elements
$1$ to
M of one of the first
MPL columns of the array
E are all zero – this corresponds to a zero row in either of the matrices
$A$ or
$C$.
7 Accuracy
The method is stable.
The effect of $m$ and $n$ on the time and on the number of iterations varies from problem to problem, but typically the number of iterations is a small multiple of $n$ and the total time taken is approximately proportional to $m{n}^{2}$.
Linear dependencies among the rows or columns of
$A$ and
$C$ are not necessarily a problem to the algorithm. Solutions can be obtained from rankdeficient
$A$ and
$C$. However, the algorithm requires that at every step the currently active columns of
E form a linearly independent set. If this is not the case at any step, small, random perturbations of the order of rounding error are added to the appropriate columns of
E. Normally this perturbation process will not affect the solution significantly. It does mean, however, that results may not be exactly reproducible.
9 Example
Suppose we wish to approximate in
$\left[0,1\right]$ a set of data by a curve of the form
which has nonnegative slope at the data points. Given points
$\left({t}_{i},{y}_{i}\right)$ we may form the equations
for
$\mathit{i}=1,2,\dots ,6$, for the
$6$ data points. The requirement of a nonnegative slope at the data points demands
for each
${t}_{i}$ and these form the constraints.
(Note that, for fitting with polynomials, it would usually be advisable to work with the polynomial expressed in Chebyshev series form (see the
E02 Chapter Introduction). The power series form is used here for simplicity of exposition.)
9.1 Program Text
Program Text (e02gbfe.f90)
9.2 Program Data
Program Data (e02gbfe.d)
9.3 Program Results
Program Results (e02gbfe.r)