lsqnonlin question

I have a function file called commands.m which gives the necessary inputs to another function file called fit_simp. lsqnonlin is called inside fit_simp.

My commands function file is as follows;

X=xlsread('MR01.xls',8,'AA63:AA133');
Y=xlsread('MR01.xls',8,'W63:W133');
X0=[800 1537 0.1722 7.169e-6 1];
lb = [800;0;0;0;0.7];
ub=[2000;2000;10;10;1];
StartAt = [800;1537;0.1722;7.169e-6;0.0001];
x=lsqnonlin(@(X0)fit_simp(X0,X,Y),StartAt,lb,ub);

And my fit_simp file is as follows;

function diff = fit_simp(x,X,Y)

% This function is called by lsqnonlin.
% x is a vector which contains the coefficients of the
% equation.  X and Y are the option data sets that were
% passed to lsqnonlin.

A=x(1);
B=x(2);
n=x(3);
C=x(4);
m=x(5);

[total_readings,epsilon_dot_QS,epsilon_dot_MR,TM,TR,rho,Cp] = GetMRDetails;

for i=1:total_readings

      if (i~=1)

          d_epsilon=(X(i)-X(i-1));
          sigma=diff(i-1)-Y(i-1);
          dT=abs((1/(rho*Cp))*(sigma*d_epsilon));
          TH=dT/(TM-TR);

          diff(i)=(A+B*(X(i)^n)+C*log(epsilon_dot_QS/epsilon_dot_MR)+(1-(TH)^m));

      else

          diff(i)=(A+B*(X(i)^n)+C*log(epsilon_dot_QS/epsilon_dot_MR));

      end
  end

When I call commands by typing_ >>commands_ in the commands window, I get the following message:

Maximum number of function evaluations exceeded; increase options.MaxFunEvals

But when I type options in the commands window, it tells me;

??? Undefined function or variable 'options'.

Pleae can anyone guide me what is going wrong? How to increase the value of options.MaxFunEvals?Please help.....

Answer 1

[the cyclist]

You have to use the optimset() function to determine the options that are being referred to here.

[Shalini]

I called optimset but it gave me the following;

>> optimset

Display: [ off | iter | iter-detailed | notify | notify-detailed | final | final-detailed ]

MaxFunEvals: [ positive scalar ]

MaxIter: [ positive scalar ]

TolFun: [ positive scalar ]

TolX: [ positive scalar ]

FunValCheck: [ on | {off} ]

OutputFcn: [ function | {[]} ]

PlotFcns: [ function | {[]} ]

Algorithm: [ active-set | interior-point | levenberg-marquardt | trust-region-dogleg | trust-region-reflective ]

AlwaysHonorConstraints: [ none | {bounds} ]

BranchStrategy: [ mininfeas | {maxinfeas} ]

DerivativeCheck: [ on | {off} ]

Diagnostics: [ on | {off} ]

DiffMaxChange: [ positive scalar | {1e-1} ]

DiffMinChange: [ positive scalar | {1e-8} ]

FinDiffType: [ {forward} | central ]

GoalsExactAchieve: [ positive scalar | {0} ]

GradConstr: [ on | {off} ]

GradObj: [ on | {off} ]

HessFcn: [ function | {[]} ]

Hessian: [ user-supplied | bfgs | lbfgs | fin-diff-grads | on | off ]

HessMult: [ function | {[]} ]

HessPattern: [ sparse matrix | {sparse(ones(numberOfVariables))} ]

HessUpdate: [ dfp | steepdesc | {bfgs} ]

InitBarrierParam: [ positive scalar | {0.1} ]

InitialHessType: [ identity | {scaled-identity} | user-supplied ]

InitialHessMatrix: [ scalar | vector | {[]} ]

InitTrustRegionRadius: [ positive scalar | {sqrt(numberOfVariables)} ]

Jacobian: [ on | {off} ]

JacobMult: [ function | {[]} ]

JacobPattern: [ sparse matrix | {sparse(ones(Jrows,Jcols))} ]

LargeScale: [ on | off ]

LevenbergMarquardt: [ {on} | off ]

LineSearchType: [ cubicpoly | {quadcubic} ]

MaxNodes: [ positive scalar | {1000*numberOfVariables} ]

MaxPCGIter: [ positive scalar | {max(1,floor(numberOfVariables/2))} ]

MaxProjCGIter: [ positive scalar | {2*(numberOfVariables-numberOfEqualities)} ]

MaxRLPIter: [ positive scalar | {100*numberOfVariables} ]

MaxSQPIter: [ positive scalar | {10*max(numberOfVariables,numberOfInequalities+numberOfBounds)} ]

MaxTime: [ positive scalar | {7200} ]

MeritFunction: [ singleobj | {multiobj} ]

MinAbsMax: [ positive scalar | {0} ]

NodeDisplayInterval: [ positive scalar | {20} ]

NodeSearchStrategy: [ df | {bn} ]

NonlEqnAlgorithm: [ {dogleg} | lm | gn ]

ObjectiveLimit: [ scalar | {-1e20} ]

PrecondBandWidth: [ positive scalar | 0 | Inf ]

RelLineSrchBnd: [ positive scalar | {[]} ]

RelLineSrchBndDuration: [ positive scalar | {1} ]

ScaleProblem: [ none | obj-and-constr | jacobian ]

Simplex: [ on | {off} ]

SubproblemAlgorithm: [ cg | {ldl-factorization} ]

TolCon: [ positive scalar ]

TolConSQP: [ positive scalar | {1e-6} ]

TolPCG: [ positive scalar | {0.1} ]

TolProjCG: [ positive scalar | {1e-2} ]

TolProjCGAbs: [ positive scalar | {1e-10} ]

TolRLPFun: [ positive scalar | {1e-6} ]

TolXInteger: [ positive scalar | {1e-8} ]

TypicalX: [ vector | {ones(numberOfVariables,1)} ]

UseParallel: [ always | {never} ]

>> optimset.MaxFunEvals

Display: [ off | iter | iter-detailed | notify | notify-detailed | final | final-detailed ]

MaxFunEvals: [ positive scalar ]

MaxIter: [ positive scalar ]

TolFun: [ positive scalar ]

TolX: [ positive scalar ]

FunValCheck: [ on | {off} ]

OutputFcn: [ function | {[]} ]

PlotFcns: [ function | {[]} ]

Algorithm: [ active-set | interior-point | levenberg-marquardt | trust-region-dogleg | trust-region-reflective ]

AlwaysHonorConstraints: [ none | {bounds} ]

BranchStrategy: [ mininfeas | {maxinfeas} ]

DerivativeCheck: [ on | {off} ]

Diagnostics: [ on | {off} ]

DiffMaxChange: [ positive scalar | {1e-1} ]

DiffMinChange: [ positive scalar | {1e-8} ]

FinDiffType: [ {forward} | central ]

GoalsExactAchieve: [ positive scalar | {0} ]

GradConstr: [ on | {off} ]

GradObj: [ on | {off} ]

HessFcn: [ function | {[]} ]

Hessian: [ user-supplied | bfgs | lbfgs | fin-diff-grads | on | off ]

HessMult: [ function | {[]} ]

HessPattern: [ sparse matrix | {sparse(ones(numberOfVariables))} ]

HessUpdate: [ dfp | steepdesc | {bfgs} ]

InitBarrierParam: [ positive scalar | {0.1} ]

InitialHessType: [ identity | {scaled-identity} | user-supplied ]

InitialHessMatrix: [ scalar | vector | {[]} ]

InitTrustRegionRadius: [ positive scalar | {sqrt(numberOfVariables)} ]

Jacobian: [ on | {off} ]

JacobMult: [ function | {[]} ]

JacobPattern: [ sparse matrix | {sparse(ones(Jrows,Jcols))} ]

LargeScale: [ on | off ]

LevenbergMarquardt: [ {on} | off ]

LineSearchType: [ cubicpoly | {quadcubic} ]

MaxNodes: [ positive scalar | {1000*numberOfVariables} ]

MaxPCGIter: [ positive scalar | {max(1,floor(numberOfVariables/2))} ]

MaxProjCGIter: [ positive scalar | {2*(numberOfVariables-numberOfEqualities)} ]

MaxRLPIter: [ positive scalar | {100*numberOfVariables} ]

MaxSQPIter: [ positive scalar | {10*max(numberOfVariables,numberOfInequalities+numberOfBounds)} ]

MaxTime: [ positive scalar | {7200} ]

MeritFunction: [ singleobj | {multiobj} ]

MinAbsMax: [ positive scalar | {0} ]

NodeDisplayInterval: [ positive scalar | {20} ]

NodeSearchStrategy: [ df | {bn} ]

NonlEqnAlgorithm: [ {dogleg} | lm | gn ]

ObjectiveLimit: [ scalar | {-1e20} ]

PrecondBandWidth: [ positive scalar | 0 | Inf ]

RelLineSrchBnd: [ positive scalar | {[]} ]

RelLineSrchBndDuration: [ positive scalar | {1} ]

ScaleProblem: [ none | obj-and-constr | jacobian ]

Simplex: [ on | {off} ]

SubproblemAlgorithm: [ cg | {ldl-factorization} ]

TolCon: [ positive scalar ]

TolConSQP: [ positive scalar | {1e-6} ]

TolPCG: [ positive scalar | {0.1} ]

TolProjCG: [ positive scalar | {1e-2} ]

TolProjCGAbs: [ positive scalar | {1e-10} ]

TolRLPFun: [ positive scalar | {1e-6} ]

TolXInteger: [ positive scalar | {1e-8} ]

TypicalX: [ vector | {ones(numberOfVariables,1)} ]

UseParallel: [ always | {never} ]

How to change the default value of MaxFunEvals?

Answer 2

[Shalini]

Thanks..Done the follwoing the commands functiona nd then it worked;

options = optimset('MaxFunEvals',10000); x=lsqnonlin(@(X0)fit_simp(X0,X,Y),StartAt,lb,ub,options);