Package nleqslv

February 5, 2014
Type Package Title Solve systems of non linear equations Version 2.1.1 Date 2013-11-25 Author Berend Hasselman Maintainer Berend Hasselman <bhh@xs4all.nl> Description Solve a system of non linear equations using a Broyden or a Newton method with a choice of global strategies such as line search and trust region. There are options for using a numerical or user supplied Jacobian and for specifying that a numerical Jacobian is banded. License GPL (>= 2) NeedsCompilation yes Repository CRAN Date/Publication 2014-02-05 20:48:01

R topics documented:
nleqlsv-package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . nleqslv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . nleqslv-iterationreport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Index 2 2 8 11

nleqslv

nleqlsv-package

Solving Nonlinear Systems of Equations

Description The nleqlsv package provides two algorithms for solving (dense) non linear systems of equations. The methods provided are a Broyden Secant method where the matrix of derivatives is updated after each major iteration using the Broyden rank 1 update. a full Newton method where the Jacobian matrix of derivatives is recalculated at each iteration Both methods utilise global strategies such as line search or trust region methods whenever the standard Newton/Broyden step does not lead to a point closer to a root of the equation system. Line search may be either quadratic or geometric. The trust region methods are either the double dogleg or the Powell single dogleg method. There is a facility for specifying that the jacobian is banded; this can signicantly speedup the calculation of a numerical Jacobian when the number of sub- and super diagonals is small compared to the size of the system of equations. For example the Jacobian of a tridiagonal system can be calculated with only three evaluations of the function. The algorithms provided in this package are derived from Dennis and Schnabel (1996). The code is written in Fortran 77 and Fortran 95 and uses Lapack and BLAS routines as provided by the R system. Author(s) Berend Hasselman <bhh@xs4all.nl> References Dennis, J.E. Jr and Schnabel, R.B. (1996), Numerical Methods for Unconstrained Optimization and Nonlinear Equations, Siam.

nleqslv

Solving systems of non linear equations with Broyden or Newton

Description The function solves a system of non linear equations with either a Broyden or a full Newton method. It provides line search and trust region global strategies for difcult systems.

nleqslv Usage nleqslv(x, fn, jac=NULL,..., method = c("Broyden", "Newton"), global = c("dbldog", "pwldog", "qline", "gline", "none"), xscalm = c("fixed","auto"), jacobian=FALSE, control = list() ) Arguments x fn jac ... method global xscalm jacobian control Details The algorithms implemented in nleqslv are based on Dennis and Schnabel (1996). A list or numeric vector of starting estimates. A function of x returning the function values.

A function to return the Jacobian for the fn function. If not supplied numerical derivatives will be used. Further arguments to be passed to fn and jac. The method used for nding a solution. See Details. The global strategy to apply. See Details. The type of x scaling to use. See Details. A logical indicating if the estimated (approximate) jacobian in the solution should be returned. See Details. A list of control parameters. See Details.

Method Broyden starts with a computed Jacobian of the function and then updates this Jacobian after each successful iteration using the so-called Broyden update. This method often shows super linear convergence towards a solution. When nleqslv determines that it cannot continue with the current Broyden matrix it will compute a new Jacobian. Method Newton calculates a Jacobian of the function fn at each iteration. Close to a solution this method will usually show quadratic convergence. Both methods apply a so-called (backtracking) global strategy to nd a better (more acceptable) iterate. The function criterion used by the algorithm is the sum of squares of the function values and acceptable means sufcient decrease of the current function criterion value compared to that of the previous iteration. A comprehensive discussion of these issues can be found in Dennis and Schnabel (1996). Both methods apply a QR-decomposition to the Jacobian as implemented in Lapack. The Broyden method applies a rank-1 update to the Jacobian at the end of each iteration and uses a modied version of the algorithm described in Dennis and Schnabel (1996). The global argument indicates which global strategy to use or to use no global strategy qline a quadratic line search gline a geometric line search dbldog a trust region method using the double dogleg method as described in Dennis and Schnabel (1996)

nleqslv pwldog a trust region method using the Powell dogleg method as developed by Powell (1970). none Only a pure local Newton or Broyden iteration is used. The maximum stepsize (see below) is taken into account. The default maximum number of iterations (see below) is set to 20. The double dogleg method is the default global strategy employed by this package. The parameters x may be scaled during the search for a zero of fn. The xscalm argument provides two possibilities for scaling fixed the scaling factors are set to the values supplied in the control argument and remain unchanged during the iterations. The scaling factor of any element of x should be set to the inverse of the typical value of that element of x, ensuring that all elements of x are approximately equal in size. auto the scaling factors are calculated from the euclidean norms of the columns of the Jacobian matrix. When a new Jacobian is computed, the scaling values will be set to the euclidean norm of the corresponding column if that is larger than the current scaling value. Thus the scaling values will not decrease during the iteration. This is the method described in Mor (1978). Usually manual scaling is preferable. The algorithm cannot cope with a singular or ill-conditioned Jacobian. When a Jacobian becomes singular or very ill-conditioned the algorithm will return an error. A Jacobian is considered to be very ill-conditioned when the estimated inverse condition is less than or equal to a specied tolerance with a default value equal to 1012 . When the function to be solved returns non-nite function values for a parameter vector x and the algorithm is not evaluating a numerical Jacobian, then any non-nite values will be replaced by a large number forcing the algorithm to backtrack, i.e. decrease the line search factor or decrease the trust region radius. When the algorithm is evaluating a numerical Jacobian, it will stop with an error message on detecting non-nite function values. An error message will also be issued when a user supplied jacobian contains non-nite entries. When the jacobian argument is set to TRUE the nal Jacobian or Broyden matrix will be returned in the return list. The default value is FALSE; i.e. to not return the nal matrix. There is no guarantee that the nal Broyden matrix resembles the actual Jacobian. The control argument is a list that can supply any of the following components: xtol The relative steplength tolerance. When the relative steplength of all scaled x values is smaller than this value convergence is declared. The default value is 108 . ftol The function value tolerance. Convergence is declared when the largest absolute function value is smaller than ftol. The default value is 108 . btol The backtracking tolerance. When the relative steplength in a backtracking step to nd an acceptable point is smaller than the backtracking tolerance, the backtracking is terminated. In the Broyden method a new Jacobian will be calculated if the Jacobian is outdated. The default value is 103 . cndtol The tolerance of the test for ill conditioning of the Jacobian or Broyden approximation. If less than the machine precision it will be silently set to the machine precision. When the estimated inverse condition of the (approximated) Jacobian matrix is less than or equal to the value of cndtol the matrix is deemed to be ill-conditioned. The default value is 1012 . sigma Reduction factor for the geometric line search. The default value is 0.5.

nleqslv

scalex a vector of scaling values for the parameters. The inverse of a scale value is an indication of the size of a parameter. The default value is 1.0 for all scale values. maxit The maximum number of major iterations. The default value is 150 if a global strategy has been specied. If no global strategy has been specied the default is 20. trace Non-negative integer. A value of 1 will give a detailed report of the progress of the iteration. For a description see Iteration report. chkjac A logical value indicating whether to check a user supplied Jacobian, if supplied. The default value is FALSE. The rst 10 errors are printed. The code for this check is derived from the code in Bouaricha and Schnabel (1997). delta Initial (scaled) trust region radius. A value of 1.0 is replaced by the length of the Cauchy step in the initial point. A value of 2.0 is replaced by the length of the Newton step in the initial point. Any value less than or equal to 0 and not equal to 2.0, will be replaced by 1.0; the algorithm will thus start with the length of the Cauchy step in the initial point. If it is positive it will be set to the smaller of the value supplied or the maximum stepsize. The default is 2.0. stepmax Maximum scaled stepsize. If this is negative then the maximum stepsize is set to the largest positive representable number. The default is 1.0, so there is no default maximum stepsize. dsub Number of non zero subdiagonals of a banded Jacobian. The default is to assume that the Jacobian is not banded. Must be specied if dsuper has been specied and must be larger than zero when dsuper is zero. dsuper Number of non zero super diagonals of a banded Jacobian. The default is to assume that the Jacobian is not banded. Must be specied if dsub has been specied and must be larger than zero when dsub is zero. Value A list containing components x fvec termcd nal values for x function values termination code as integer. The values returned are 1 Function criterion is near zero. Convergence of function values has been achieved. 2 x-values within tolerance. This means that the relative distance between two consecutive x-values is smaller than xtol. 3 No better point found. This means that the algorithm has stalled and cannot nd an acceptable new point. This may or may not indicate acceptably small function values. 4 Iteration limit maxit exceeded. 5 Jacobian is too ill-conditioned. 6 Jacobian is singular. -10 User supplied Jacobian is most likely incorrect. message a string describing the termination code

6 scalex njcnt nfcnt iter

nleqslv a vector containing the scaling factors, which will be the nal values when automatic scaling was selected number of Jacobian evaluations number of function evaluations, excluding those required for calculating a Jacobian and excluding the initial function evaluation (at iteration 0) number of outer iterations used by the algorithm. This excludes the initial iteration. The number of backtracks can be calculated as the difference between items nfcnt and iter of the outputlist. the nal Jacobian or the Broyden approximation if jacobian was set to TRUE. If no iterations were executed, as may happen when the initial starting values are sufciently close the solution, there is no Broyden approximation and the returned matrix will always be the actual Jacobian. If the matrix is singular or too ill-conditioned the returned matrix is of no value.

jac

Warning You cannot use this function recursively. Thus function fn should not in its turn call nleqslv. References Bouaricha, A. and Schnabel, R.B. (1997), Algorithm 768: TENSOLVE: A Software Package for Solving Systems of Nonlinear Equations and Nonlinear Least-squares Problems Using Tensor Methods, Transactions on Mathematical Software, 23, 2, pp. 174195. Dennis, J.E. Jr and Schnabel, R.B. (1996), Numerical Methods for Unconstrained Optimization and Nonlinear Equations, Siam. Mor, J.J. (1978), The Levenberg-Marquardt Algorithm, Implementation and Theory, In Numerical Analysis, G.A. Watson (Ed.), Lecture Notes in Mathematics 630, Springer-Verlag, pp. 105116. Golub, G.H and C.F. Van Loan (1996), Matrix Computations (3rd edition), The John Hopkins University Press. Powell, M.J.D. (1970), A hybrid method for nonlinear algebraic equations, In Numerical Methods for Nonlinear Algebraic Equations, P. Rabinowitz (Ed.), Gordon \& Breach. Powell, M.J.D. (1970), A Fortran subroutine for solving systems nonlinear equations, In Numerical Methods for Nonlinear Algebraic Equations, P. Rabinowitz (Ed.), Gordon \& Breach. Reichel, L. and W.B. Gragg (1990), Algorithm 686: FORTRAN subroutines for updating the QR decomposition, ACM Trans. Math. Softw., 16, 4, pp. 369377. Examples
# Dennis Schnabel example 6.5.1 page 149 dslnex <- function(x) { y <- numeric(2) y[1] <- x[1]^2 + x[2]^2 - 2 y[2] <- exp(x[1]-1) + x[2]^3 - 2 y } jacdsln <- function(x) {

nleqslv
n <- length(x) Df <- matrix(numeric(n*n),n,n) Df[1,1] <- 2*x[1] Df[1,2] <- 2*x[2] Df[2,1] <- exp(x[1]-1) Df[2,2] <- 3*x[2]^2 Df } BADjacdsln <- function(x) { n <- length(x) Df <- matrix(numeric(n*n),n,n) Df[1,1] <- 4*x[1] Df[1,2] <- 2*x[2] Df[2,1] <- exp(x[1]-1) Df[2,2] <- 5*x[2]^2 Df } xstart <- c(2,0.5) fstart <- dslnex(xstart) xstart fstart # a solution is c(1,1) nleqslv(xstart, dslnex, control=list(btol=.01)) # Cauchy start nleqslv(xstart, dslnex, control=list(trace=1,btol=.01,delta=-1.0)) # Newton start nleqslv(xstart, dslnex, control=list(trace=1,btol=.01,delta=-2.0)) # final Broyden approximation of Jacobian (quite good) z <- nleqslv(xstart, dslnex, jacobian=TRUE,control=list(btol=.01)) z$x z$jac jacdsln(z$x) # different initial start; not a very good final approximation xstart <- c(0.5,2) z <- nleqslv(xstart, dslnex, jacobian=TRUE,control=list(btol=.01)) z$x z$jac jacdsln(z$x)

## Not run: # no global strategy but limit stepsize # but look carefully: a different solution is found nleqslv(xstart, dslnex, method="Newton", global="none", control=list(trace=1,stepmax=5))

nleqslv-iterationreport

# but if the stepsize is limited even more the c(1,1) solution is found nleqslv(xstart, dslnex, method="Newton", global="none", control=list(trace=1,stepmax=2)) # Broyden also finds the c(1,1) solution when the stepsize is limited nleqslv(xstart, dslnex, jacdsln, method="Broyden", global="none", control=list(trace=1,stepmax=2)) ## End(Not run)

nleqslv-iterationreport Detailed iteration report of nleqslv

Description The format of the iteration report provided by nleqslv when the trace component of the control parameter has been set to 1. Details All iteration reports consist of a series of columns with a header summarising the contents. Common column headers are Iter Iteration counter Jac Jacobian type. The Jacobian type is indicated by N for a Newton Jacobian or B for a Broyden updated matrix; optionally followed by the letter s indicating a totally singular matrix or the letter i indicating an ill-conditioned matrix. Unless the Jacobian is singular, the Jacobian type is followed by an estimate of the inverse condition number of the Jacobian in parentheses as computed by Lapack. This column will be blank when backtracking is active. Fnorm square of the euclidean norm of function values / 2 Largest |f| innity norm of f (x) at the current point A sample iteration report for qline is (some intercolumn space has been removed to make the table t) Iter Jac Lambda Ftarg Fnorm 0 2.886812e+00 1 N(9.6e-03) 1.0000 2.886235e+00 5.787362e+05 1 0.1000 2.886754e+00 9.857947e+00 1 0.0100 2.886806e+00 2.866321e+00 2 B(2.2e-02) 1.0000 2.865748e+00 4.541965e+03 Largest |f| 2.250000e+00 1.070841e+03 3.214799e+00 2.237878e+00 9.341610e+01

The column headed by Lambda shows the value of the line search parameter. The column headed by Ftarg follows from a sufcient decrease requirement and is the value below which Fnorm must drop if the current step is to be accepted. The iteration report for none is almost the same as the above report, except that the column labelled Ftarg is omitted. The column Lambda gives the ratio of the maximum stepsize and the length of the step taken by the algorithm. It is either 1.0 exactly or smaller.

nleqslv-iterationreport

A sample iteration report for dbldog is (some intercolumn space has been removed to make the table t) Iter 0 1 1 2 3 3 4 5 6 7 8 9 10 11 Jac N(9.6e-03) C W B(1.1e-02) W B(7.3e-02) W C B(8.3e-02) W B(1.8e-01) N B(1.8e-01) N B(1.9e-01) N B(1.6e-01) N B(1.5e-01) N B(1.5e-01) N B(1.5e-01) N Lambda Eta Dlt0 Dltn Fnorm 2.886812e+00 1.699715e-01 1.699715e-01 1.277667e-01 5.067893e-01 5.440250e-02 3.576547e-02 6.566182e-03 4.921645e-04 4.960629e-06 1.545503e-08 2.968676e-11 4.741792e-14 6.451792e-19 Largest |f| 2.250000e+00 5.421673e-01 5.421673e-01 5.043571e-01 7.973542e-01 2.726084e-01 2.657553e-01 8.555110e-02 3.094104e-02 2.826064e-03 1.757498e-04 5.983765e-06 2.198380e-07 8.118586e-10

0.9544 0.4671 0.9343 0.0833 0.9544 0.9343 0.4671 0.1154 0.4851 0.4671 0.4671 0.7879 0.7289 0.4671 0.0759 0.7289 0.0759 0.1519 0.5307 0.3271 0.1519 0.3037 0.6674 0.2191 0.4383 0.9801 0.0376 0.0752 0.7981 0.0157 0.0313 0.3942 0.0029 0.0058 0.6536 0.0001 0.0003 0.4730 0.0000 0.0000 0.9787 0.0000 0.0000

After the column for the Jacobian the letters indicate the following C a fraction ( 1.0) of the Cauchy or steepest descent step is taken where the fraction is the ratio of the trust region radius and the Cauchy steplength. W a convex combination of the Cauchy and eta*(Newton step) is taken. The number in the column headed by Lambda is the weight of the partial Newton step. P a fraction ( 1.0) of the partial Newton step, equal to eta*(Newton step), is taken where the fraction is the ratio of the trust region radius and the partial Newton steplength. N a normal full Newton step is taken. The number in the column headed by Eta is calculated from an upper limit on the ratio of the length of the steepest descent direction and the length of the Newton step. See Dennis and Schnabel (1996) pp.139ff for the details. The column headed by Dlt0 gives the size of the trust region at the start of the current iteration. The column headed by Dltn gives the size of the trust region when the current step has been accepted by the dogleg algorithm. The size of the trust region is decreased when the actual reduction of the function value norm does not agree well with the predicted reduction from the linear approximation of the function. And increased when the actual and predicted reduction are sufciently close. Normally the initial trust region radius is the same as the nal trust region radius of the previous iteration but the size of the trust region is restricted by the size of the current Newton step. So when full Newton steps are being taken, the trust region radius at the start of an iteration may be less than the nal value of the previous iteration. A sample iteration report for pwldog is (some intercolumn space has been removed to make the table t) Iter Jac Lambda Dlt0 Dltn Fnorm Largest |f| 0 2.886812e+00 2.250000e+00 1 N(9.6e-03) C 0.4671 0.9343 1.699715e-01 5.421673e-01 1 W 0.0794 0.9343 0.4671 1.699715e-01 5.421673e-01

10 2 3 3 4 4 5 B(1.1e-02) W 0.0559 B(7.3e-02) W 0.5662 W 0.0237 B(8.8e-02) W 0.2306 W 0.4769 B(1.9e-01) N 0.4671 0.4671 0.0960 0.1921 0.3842 0.1375 0.4671 0.0960 0.1921 0.3842 0.1921 0.2750 1.205661e-01 4.119560e-01 4.426507e-02 2.303135e-02 2.303135e-02 8.014508e-04

nleqslv-iterationreport 4.890487e-01 7.254441e-01 2.139252e-01 2.143943e-01 2.143943e-01 3.681498e-02

This is much simpler than the double dogleg report, since the (single) dogleg takes either a steepest descent step, a convex combination of the steepest descent and Newton steps or a full Newton step. The number in the column Lambda is the weight of the Newton step. It is a special case of the double dogleg method with eta equal to 1. See Also nleqslv

Index
Topic nonlinear nleqslv, 2 Topic optimize nleqslv, 2 Topic package nleqlsv-package, 2 Iteration report (nleqslv-iterationreport), 8 nleqlsv-package, 2 nleqlsv.Intro (nleqlsv-package), 2 nleqslv, 2, 10 nleqslv-iterationreport, 8

11