The routine may be called by the names e02gaf or nagf_fit_glin_l1sol.
Given a matrix with rows and columns and a vector with elements, the routine calculates an solution to the overdetermined system of equations
That is to say, it calculates a vector , with elements, which minimizes the norm (the sum of the absolute values) of the residuals
where the residuals are given by
Here is the element in row and column of , is the th element of and the th element of . The matrix need not be of full rank.
Typically in applications to data fitting, data consisting of points with coordinates are to be approximated in the norm by a linear combination of known functions ,
This is equivalent to fitting an solution to the overdetermined system of equations
Thus if, for each value of and , the element of the matrix in the previous paragraph is set equal to the value of and is set equal to , the solution vector will contain the required values of the . Note that the independent variable above can, instead, be a vector of several independent variables (this includes the case where each is a function of a different variable, or set of variables).
The algorithm is a modification of the simplex method of linear programming applied to the primal formulation of the problem (see Barrodale and Roberts (1973) and Barrodale and Roberts (1974)). The modification allows several neighbouring simplex vertices to be passed through in a single iteration, providing a substantial improvement in efficiency.
Barrodale I and Roberts F D K (1973) An improved algorithm for discrete linear approximation SIAM J. Numer. Anal.10 839–848
Barrodale I and Roberts F D K (1974) Solution of an overdetermined system of equations in the -norm Comm. ACM17(6) 319–320
1: – IntegerInput
On entry: the number of equations, (the number of rows of the matrix ).
2: – Real (Kind=nag_wp) arrayInput/Output
On entry: must contain , the element in the th row and th column of the matrix , for and . The remaining elements need not be set.
On exit: contains the last simplex tableau generated by the simplex method.
3: – IntegerInput
On entry: the first dimension of the array a as declared in the (sub)program from which e02gaf is called.
4: – Real (Kind=nag_wp) arrayInput/Output
On entry: must contain , the th element of the vector , for .
On exit: the
th residual corresponding to the solution vector , for .
5: – IntegerInput
On entry: , where is the number of unknowns (the number of columns of the matrix ).
6: – Real (Kind=nag_wp)Input
On entry: a non-negative value. In general toler specifies a threshold below which numbers are regarded as zero. The recommended threshold value is where is the machine precision. The recommended value can be computed within the routine by setting toler to zero. If premature termination occurs a larger value for toler may result in a valid solution.
7: – Real (Kind=nag_wp) arrayOutput
On exit: contains the th element of the solution vector , for . The elements and are unused.
8: – Real (Kind=nag_wp)Output
On exit: the sum of the absolute values of the residuals for the solution vector .
9: – IntegerOutput
On exit: the computed rank of the matrix .
10: – IntegerOutput
On exit: the number of iterations taken by the simplex method.
11: – Integer arrayWorkspace
12: – IntegerInput/Output
On entry: ifail must be set to , or to set behaviour on detection of an error; these values have no effect when no error is detected.
A value of causes the printing of an error message and program execution will be halted; otherwise program execution continues. A value of means that an error message is printed while a value of means that it is not.
If halting is not appropriate, the value or is recommended. If message printing is undesirable, then the value is recommended. Otherwise, the value is recommended. When the value or is used it is essential to test the value of ifail on exit.
On exit: unless the routine detects an error or a warning has been flagged (see Section 6).
6Error Indicators and Warnings
If on entry or , explanatory error messages are output on the current error message unit (as defined by x04aaf).
Errors or warnings detected by the routine:
An optimal solution has been obtained, but may not be unique.
Premature termination due to rounding errors. Try using larger value of toler: .
On entry, and .
On entry, .
On entry, and .
More than iterations were performed. e02gaf has terminated without calculating a solution. The output data from the routine is as computed on the last good iteration. Consider increasing the value of toler. Alternatively, may be ill conditioned—try scaling its columns.
An unexpected error has been triggered by this routine. Please
See Section 7 in the Introduction to the NAG Library FL Interface for further information.
Your licence key may have expired or may not have been installed correctly.
See Section 8 in the Introduction to the NAG Library FL Interface for further information.
Dynamic memory allocation failed.
See Section 9 in the Introduction to the NAG Library FL Interface for further information.
Experience suggests that the computational accuracy of the solution is comparable with the accuracy that could be obtained by applying Gaussian elimination with partial pivoting to the equations satisfied by this algorithm (i.e., those equations with zero residuals). The accuracy, therefore, varies with the conditioning of the problem, but has been found generally very satisfactory in practice.
8Parallelism and Performance
e02gaf is not threaded in any implementation.
The effects of and on the time and on the number of iterations in the Simplex Method vary from problem to problem, but typically the number of iterations is a small multiple of and the total time taken is approximately proportional to .
It is recommended that, before the routine is entered, the columns of the matrix are scaled so that the largest element in each column is of the order of unity. This should improve the conditioning of the matrix, and also enable the argument toler to perform its correct function. The solution obtained will then, of course, relate to the scaled form of the matrix. Thus if the scaling is such that, for each , the elements of the th column are multiplied by the constant , the element of the solution vector must be multiplied by if it is desired to recover the solution corresponding to the original matrix .
Suppose we wish to approximate a set of data by a curve of the form
where , and are unknown. Given values at points we may form the overdetermined set of equations for , and