lsqnonlin
Solve nonlinear least-squares (nonlinear data-fitting) problems
Syntax
Description
Nonlinear least-squares solver
Solves nonlinear least-squares curve fitting problems of the form
subject to the constraints
x,lb
, andub
can be vectors or matrices; seeMatrix Arguments.
Do not specify the objective function as the scalar value
(the sum of squares).lsqnonlin
requires the objective function to be thevector-valued function
starts at the pointx
= lsqnonlin(fun
,x0
)x0
and finds a minimum of the sum of squares of the functions described infun
. The functionfun
should return a vector (or array) of values and not the sum of squares of the values. (The algorithm implicitly computes the sum of squares of the components offun(x)
.)
Note
Passing Extra Parametersexplains how to pass extra parameters to the vector functionfun(x)
, if necessary.
defines a set of lower and upper bounds on the design variables inx
= lsqnonlin(fun
,x0
,lb
,ub
)x
, so that the solution is always in the rangelb
≤x
≤ub
. You can fix the solution componentx(i)
by specifyinglb(i) = ub(i)
.
Note
If the specified input bounds for a problem are inconsistent, the outputx
isx0
and the outputsresnorm
andresidual
are[]
.
Components ofx0
that violate the boundslb ≤ x ≤ ub
are reset to the interior of the box defined by the bounds. Components that respect the bounds are not changed.
Examples
Fit a Simple Exponential
Fit a simple exponential decay curve to data.
Generate data from an exponential decay model plus noise. The model is
with ranging from 0 through 3, and normally distributed noise with mean 0 and standard deviation 0.05.
rngdefault% for reproducibilityd = linspace(0,3); y = exp(-1.3*d) + 0.05*randn(size(d));
The problem is: given the data (d
,y
), find the exponential decay rate that best fits the data.
Create an anonymous function that takes a value of the exponential decay rate and returns a vector of differences from the model with that decay rate and the data.
fun = @(r)exp(-d*r)-y;
Find the value of the optimal decay rate. Arbitrarily choose an initial guessx0
= 4.
x0 = 4; x = lsqnonlin(fun,x0)
Local minimum possible. lsqnonlin stopped because the final change in the sum of squares relative to its initial value is less than the value of the function tolerance.
x = 1.2645
Plot the data and the best-fitting exponential curve.
plot(d,y,'ko',d,exp(-x*d),'b-') legend('Data','Best fit') xlabel('t') ylabel('exp(-tx)')
Fit a Problem with Bound Constraints
Find the best-fitting model when some of the fitting parameters have bounds.
Find a centering and scaling that best fit the function
to the standard normal density,
Create a vectort
of data points, and the corresponding normal density at those points.
t = linspace(-4,4); y = 1/sqrt(2*pi)*exp(-t.^2/2);
Create a function that evaluates the difference between the centered and scaled function from the normaly
, withx(1)
as the scaling
andx(2)
as the centering
.
fun = @(x)x(1)*exp(-t).*exp(-exp(-(t-x(2)))) - y;
Find the optimal fit starting fromx0
=[1/2,0]
, with the scaling
between 1/2 and 3/2, and the centering
between -1 and 3.
磅= [1/2,1];乌兰巴托= (3/2,3);x0 = [1 /2,0]; x = lsqnonlin(fun,x0,lb,ub)
Local minimum possible. lsqnonlin stopped because the final change in the sum of squares relative to its initial value is less than the value of the function tolerance.
x =1×20.8231 -0.2444
Plot the two functions to see the quality of the fit.
plot(t,y,'r-',t,fun(x)+y,'b-') xlabel('t') legend('Normal density','Fitted function')
Least Squares with Linear Constraint
Consider the following objective function, a sum of squares:
The code for this objective function appears as themyfun
function at theend of this example.
Minimize this function subject to the linear constraint . Write this constraint as .
A = [1 -1/2]; b = 0;
Impose the bounds , , , and .
lb = [0 0]; ub = [2 4];
Start the optimization process from the pointx0 = [0.3 0.4]
.
x0 = [0.3 0.4];
The problem has no linear equality constraints.
Aeq = []; beq = [];
Run the optimization.
x = lsqnonlin(@myfun,x0,lb,ub,A,b,Aeq,beq)
Local minimum found that satisfies the constraints. Optimization completed because the objective function is non-decreasing in feasible directions, to within the value of the optimality tolerance, and constraints are satisfied to within the value of the constraint tolerance.
x =1×20.1695 0.3389
functionF = myfun(x) k = 1:10; F = 2 + 2*k - exp(k*x(1)) - 2*exp(2*k*(x(2)^2));end
Nonlinear Least Squares with Nonlinear Constraint
Consider the following objective function, a sum of squares:
The code for this objective function appears as themyfun
function at theend of this example.
Minimize this function subject to the nonlinear constraint
. The code for this nonlinear constraint function appears as thenlcon
function at theend of this example.
Impose the bounds , , , and .
lb = [0 0]; ub = [2 4];
Start the optimization process from the pointx0 = [0.3 0.4]
.
x0 = [0.3 0.4];
The problem has no linear constraints.
A = []; b = []; Aeq = []; beq = [];
Run the optimization.
x = lsqnonlin(@myfun,x0,lb,ub,A,b,Aeq,beq,@nlcon)
Local minimum possible. Constraints satisfied. fmincon stopped because the size of the current step is less than the value of the step size tolerance and constraints are satisfied to within the value of the constraint tolerance.
x =1×20.2133 0.3266
functionF = myfun(x) k = 1:10; F = 2 + 2*k - exp(k*x(1)) - 2*exp(2*k*(x(2)^2));endfunction[c,ceq] = nlcon(x) ceq = []; c = sin(x(1)) - cos(x(2));end
Nonlinear Least Squares with Nondefault Options
Compare the results of a data-fitting problem when using differentlsqnonlin
algorithms.
Suppose that you have observation time dataxdata
and observed response dataydata
, and you want to find parameters
and
to fit a model of the form
Input the observation times and responses.
xdata =...[0.9 1.5 13.8 19.8 24.1 28.2 35.2 60.3 74.6 81.3]; ydata =...[455.2 428.6 124.1 67.3 43.2 28.1 13.1 -0.4 -1.3 -1.5];
Create a simple exponential decay model. The model computes a vector of differences between predicted values and observed values.
fun = @(x)x(1)*exp(x(2)*xdata)-ydata;
Fit the model using the starting pointx0 = [100,-1]
. First, use the default'trust-region-reflective'
algorithm.
x0 = [100,-1]; options = optimoptions(@lsqnonlin,'Algorithm','trust-region-reflective'); x = lsqnonlin(fun,x0,[],[],options)
Local minimum possible. lsqnonlin stopped because the final change in the sum of squares relative to its initial value is less than the value of the function tolerance.
x =1×2498.8309 -0.1013
See if there is any difference using the'levenberg-marquardt'
algorithm.
options.Algorithm ='levenberg-marquardt';x = lsqnonlin(有趣,x0,[]、[]选项)
Local minimum possible. lsqnonlin stopped because the relative size of the current step is less than the value of the step size tolerance.
x =1×2498.8309 -0.1013
The two algorithms found the same solution. Plot the solution and the data.
plot(xdata,ydata,'ko') holdontlist = linspace(xdata(1),xdata(end)); plot(tlist,x(1)*exp(x(2)*tlist),'b-') xlabelxdataylabelydatatitle('Exponential Fit to Data') legend('Data','Exponential Fit') holdoff
Nonlinear Least Squares Solution and Residual Norm
Find the that minimizes
,
and find the value of the minimal sum of squares.
Becauselsqnonlin
assumes that the sum of squares is not explicitly formed in the user-defined function, the function passed tolsqnonlin
should instead compute the vector-valued function
,
for to (that is, should have components).
Themyfun
function, which computes the 10-component vector F, appears at theend of this example.
Find the minimizing point and the minimum value, starting at the pointx0 = [0.3,0.4]
.
x0 = [0.3,0.4]; [x,resnorm] = lsqnonlin(@myfun,x0)
Local minimum possible. lsqnonlin stopped because the size of the current step is less than the value of the step size tolerance.
x =1×20.2578 0.2578
resnorm = 124.3622
Theresnorm
output is the squared residual norm, or the sum of squares of the function values.
The following function computes the vector-valued objective function.
functionF = myfun(x) k = 1:10; F = 2 + 2*k-exp(k*x(1))-exp(k*x(2));end
Examine the Solution Process
Examine the solution process both as it occurs (by setting theDisplay
option to'iter'
) and afterward (by examining theoutput
structure).
Suppose that you have observation time dataxdata
and observed response dataydata
, and you want to find parameters
and
to fit a model of the form
Input the observation times and responses.
xdata =...[0.9 1.5 13.8 19.8 24.1 28.2 35.2 60.3 74.6 81.3]; ydata =...[455.2 428.6 124.1 67.3 43.2 28.1 13.1 -0.4 -1.3 -1.5];
Create a simple exponential decay model. The model computes a vector of differences between predicted values and observed values.
fun = @(x)x(1)*exp(x(2)*xdata)-ydata;
Fit the model using the starting pointx0 = [100,-1]
. Examine the solution process by setting theDisplay
option to'iter'
. Obtain anoutput
structure to obtain more information about the solution process.
x0 = [100,-1]; options = optimoptions('lsqnonlin','Display','iter'); [x,resnorm,residual,exitflag,output] = lsqnonlin(fun,x0,[],[],options);
Norm of First-order Iteration Func-count Resnorm step optimality 0 3 359677 2.88e+04 Objective function returned Inf; trying a new point... 1 6 359677 11.6976 2.88e+04 2 9 321395 0.5 4.97e+04 3 12 321395 1 4.97e+04 4 15 292253 0.25 7.06e+04 5 18 292253 0.5 7.06e+04 6 21 270350 0.125 1.15e+05 7 24 270350 0.25 1.15e+05 8 27 252777 0.0625 1.63e+05 9 30 252777 0.125 1.63e+05 10 33 243877 0.03125 7.48e+04 11 36 243660 0.0625 8.7e+04 12 39 243276 0.0625 2e+04 13 42 243174 0.0625 1.14e+04 14 45 242999 0.125 5.1e+03 15 48 242661 0.25 2.04e+03 16 51 241987 0.5 1.91e+03 17 54 240643 1 1.04e+03 18 57 237971 2 3.36e+03 19 60 232686 4 6.04e+03 20 63 222354 8 1.2e+04 21 66 202592 16 2.25e+04 22 69 166443 32 4.05e+04 23 72 106320 64 6.68e+04 24 75 28704.7 128 8.31e+04 25 78 89.7947 140.674 2.22e+04 26 81 9.57381 2.02599 684 27 84 9.50489 0.0619927 2.27 28 87 9.50489 0.000462261 0.0114 Local minimum possible. lsqnonlin stopped because the final change in the sum of squares relative to its initial value is less than the value of the function tolerance.
Examine the output structure to obtain more information about the solution process.
output
output =struct with fields:firstorderopt: 0.0114 iterations: 28 funcCount: 87 cgiterations: 0 algorithm: 'trust-region-reflective' stepsize: 4.6226e-04 message: 'Local minimum possible....' bestfeasible: [] constrviolation: []
For comparison, set theAlgorithm
option to'levenberg-marquardt'
.
options.Algorithm ='levenberg-marquardt';[x,resnorm,residual,exitflag,output] = lsqnonlin(fun,x0,[],[],options);
First-order Norm of Iteration Func-count Resnorm optimality Lambda step 0 3 359677 2.88e+04 0.01 Objective function returned Inf; trying a new point... 1 13 340761 3.91e+04 100000 0.280777 2 16 304661 5.97e+04 10000 0.373146 3 21 297292 6.55e+04 1e+06 0.0589933 4 24 288240 7.57e+04 100000 0.0645444 5 28 275407 1.01e+05 1e+06 0.0741266 6 31 249954 1.62e+05 100000 0.094571 7 36 245896 1.35e+05 1e+07 0.0133606 8 39 243846 7.26e+04 1e+06 0.0094431 9 42 243568 5.66e+04 100000 0.0082162 10 45 243424 1.61e+04 10000 0.00777935 11 48 243322 8.8e+03 1000 0.0673933 12 51 242408 5.1e+03 100 0.675209 13 54 233628 1.05e+04 10 6.59804 14 57 169089 8.51e+04 1 54.6992 15 60 30814.7 1.54e+05 0.1 196.939 16 63 147.496 8e+03 0.01 129.795 17 66 9.51503 117 0.001 9.96069 18 69 9.50489 0.0714 0.0001 0.080486 19 72 9.50489 5.23e-05 1e-05 5.07043e-05 Local minimum possible. lsqnonlin stopped because the relative size of the current step is less than the value of the step size tolerance.
The'levenberg-marquardt'
converged with fewer iterations, but almost as many function evaluations:
output
output =struct with fields:iterations: 19 funcCount: 72 stepsize: 5.0704e-05 cgiterations: [] firstorderopt: 5.2319e-05 algorithm: 'levenberg-marquardt' message: 'Local minimum possible....' bestfeasible: [] constrviolation: []
Input Arguments
fun
—Function whose sum of squares is minimized
function handle|name of function
Function whose sum of squares is minimized, specified as a function handle or the name of a function. For the'interior-point'
algorithm,fun
必须是一个function handle.fun
is a function that accepts an arrayx
and returns an arrayF
, the objective function evaluated atx
.
Note
The sum of squares should not be formed explicitly. Instead, your function should return a vector of function values. SeeExamples.
The functionfun
can be specified as a function handle to a file:
x = lsqnonlin(@myfun,x0)
wheremyfun
is a MATLAB®function such as
functionF = myfun(x) F =...% Compute function values at x
fun
can also be a function handle for an anonymous function.
x = lsqnonlin(@(x)sin(x.*x),x0);
lsqnonlin
passesx
to your objective function in the shape of thex0
论点. For example, ifx0
is a 5-by-3 array, thenlsqnonlin
passesx
tofun
as a 5-by-3 array.
If the Jacobian can also be computedandthe'SpecifyObjectiveGradient'
option istrue
, set by
options = optimoptions('lsqnonlin','SpecifyObjectiveGradient',true)
then the functionfun
must return a second output argument with the Jacobian valueJ
(a matrix) atx
. By checking the value ofnargout
, the function can avoid computingJ
whenfun
is called with only one output argument (in the case where the optimization algorithm only needs the value ofF
but notJ
).
function[F,J] = myfun(x) F =...% Objective function values at xifnargout > 1% Two output argumentsJ =...% Jacobian of the function evaluated at xend
Iffun
returns an array ofm
components andx
hasn
elements, wheren
is the number of elements ofx0
, the JacobianJ
is anm
-by-n
matrix whereJ(i,j)
is the partial derivative ofF(i)
with respect tox(j)
. (The JacobianJ
is the transpose of the gradient ofF
.)
Example:@(x)cos(x).*exp(-x)
Data Types:char
|function_handle
|string
x0
—Initial point
real vector|real array
Initial point, specified as a real vector or real array. Solvers use the number of elements inx0
and the size ofx0
to determine the number and size of variables thatfun
accepts.
Example:x0 = [1,2,3,4]
Data Types:double
lb
—Lower bounds
real vector|real array
Lower bounds, specified as a real vector or real array. If the number of elements inx0
is equal to the number of elements inlb
, thenlb
specifies that
x(i) >= lb(i)
for alli
.
Ifnumel(lb) < numel(x0)
, thenlb
specifies that
x(i) >= lb(i)
for1 <= i <= numel(lb)
.
Iflb
has fewer elements thanx0
, solvers issue a warning.
Example:To specify that all x components are positive, uselb = zeros(size(x0))
.
Data Types:double
ub
—Upper bounds
real vector|real array
Upper bounds, specified as a real vector or real array. If the number of elements inx0
is equal to the number of elements inub
, thenub
specifies that
x(i) <= ub(i)
for alli
.
Ifnumel(ub) < numel(x0)
, thenub
specifies that
x(i) <= ub(i)
for1 <= i <= numel(ub)
.
Ifub
has fewer elements thanx0
, solvers issue a warning.
Example:To specify that all x components are less than 1, useub = ones(size(x0))
.
Data Types:double
A
—Linear inequality constraints
real matrix
Linear inequality constraints, specified as a real matrix.A
is anM
-by-N
matrix, whereM
is the number of inequalities, andN
is the number of variables (number of elements inx0
). For large problems, passA
as a sparse matrix.
A
encodes theM
linear inequalities
A*x <= b
,
wherex
is the column vector ofN
variablesx(:)
, andb
is a column vector withM
elements.
For example, consider these inequalities:
x1+ 2x2≤ 10
3x1+ 4x2≤ 20
5x1+ 6x2≤ 30,
Specify the inequalities by entering the following constraints.
A = [1,2;3,4;5,6]; b = [10;20;30];
Example:To specify that the x components sum to 1 or less, useA = ones(1,N)
andb = 1
.
Data Types:double
b
—Linear inequality constraints
real vector
Linear inequality constraints, specified as a real vector.b
is anM
-element vector related to theA
matrix. If you passb
as a row vector, solvers internally convertb
to the column vectorb(:)
. For large problems, passb
as a sparse vector.
b
encodes theM
linear inequalities
A*x <= b
,
wherex
is the column vector ofN
variablesx(:)
, andA
is a matrix of sizeM
-by-N
.
For example, consider these inequalities:
x1+ 2x2≤ 10
3x1+ 4x2≤ 20
5x1+ 6x2≤ 30.
Specify the inequalities by entering the following constraints.
A = [1,2;3,4;5,6]; b = [10;20;30];
Example:To specify that the x components sum to 1 or less, useA = ones(1,N)
andb = 1
.
Data Types:double
Aeq
—Linear equality constraints
real matrix
线性等式约束,指定为一个真正的matrix.Aeq
is anMe
-by-N
matrix, whereMe
is the number of equalities, andN
is the number of variables (number of elements inx0
). For large problems, passAeq
as a sparse matrix.
Aeq
encodes theMe
linear equalities
Aeq*x = beq
,
wherex
is the column vector ofN
variablesx(:)
, andbeq
is a column vector withMe
elements.
For example, consider these inequalities:
x1+ 2x2+ 3x3= 10
2x1+ 4x2+x3= 20,
Specify the inequalities by entering the following constraints.
Aeq = [1,2,3;2,4,1]; beq = [10;20];
Example:To specify that the x components sum to 1, useAeq = ones(1,N)
andbeq = 1
.
Data Types:double
beq
—Linear equality constraints
real vector
Linear equality constraints, specified as a real vector.beq
is anMe
-element vector related to theAeq
matrix. If you passbeq
as a row vector, solvers internally convertbeq
to the column vectorbeq(:)
. For large problems, passbeq
as a sparse vector.
beq
encodes theMe
linear equalities
Aeq*x = beq
,
wherex
is the column vector ofN
variablesx(:)
, andAeq
is a matrix of sizeMe
-by-N
.
For example, consider these equalities:
x1+ 2x2+ 3x3= 10
2x1+ 4x2+x3= 20.
Specify the equalities by entering the following constraints.
Aeq = [1,2,3;2,4,1]; beq = [10;20];
Example:To specify that the x components sum to 1, useAeq = ones(1,N)
andbeq = 1
.
Data Types:double
nonlcon
—Nonlinear constraints
function handle
Nonlinear constraints, specified as a function handle.nonlcon
is a function that accepts a vector or arrayx
and returns two arrays,c(x)
andceq(x)
.
c(x)
is the array of nonlinear inequality constraints atx
.lsqnonlin
attempts to satisfyc(x) <= 0
for all entries ofc
.(1) ceq(x)
is the array of nonlinear equality constraints atx
.lsqnonlin
attempts to satisfyceq(x) = 0
for all entries ofceq
.(2)
For example,
x = lsqnonlin(@myfun,x0,lb,ub,A,b,Aeq,beq,@mycon,options)
wheremycon
is a MATLAB function such as
function[c,ceq] = mycon(x) c =...% Compute nonlinear inequalities at x.ceq =...% Compute nonlinear equalities at x.
SpecifyConstraintGradient
option istrue
, as set byoptions = optimoptions('','SpecifyConstraintGradient',true)
nonlcon
must also return, in the third and fourth output arguments,GC
, the Jacobian ofc(x)
, andGCeq
, the Jacobian ofceq(x)
. The JacobianG(x) of a vector functionF(x) is
GC
andGCeq
can be sparse or dense. IfGC
orGCeq
is large, with relatively few nonzero entries, save running time and memory in the'interior-point'
algorithm by representing them as sparse matrices. For more information, seeNonlinear Constraints.
Data Types:function_handle
options
—Optimization options
output ofoptimoptions
|structure asoptimset
returns
Optimization options, specified as the output ofoptimoptions
or a structure asoptimset
returns.
Some options apply to all algorithms, and others are relevant for particular algorithms. SeeOptimization Options Referencefor detailed information.
Some options are absent from theoptimoptions
display. These options appear in italics in the following table. For details, seeView Optimization Options.
All Algorithms | |
|
Choose between The The For more information on choosing the algorithm, seeChoosing the Algorithm. |
CheckGradients |
Compare user-supplied derivatives (gradients of objective or constraints) to finite-differencing derivatives. Choices are For |
Diagnostics | Display diagnostic information about the function to be minimized or solved. Choices are |
DiffMaxChange | Maximum change in variables for finite-difference gradients (a positive scalar). The default is |
DiffMinChange | Minimum change in variables for finite-difference gradients (a positive scalar). The default is |
|
Level of display (seeIterative Display):
|
FiniteDifferenceStepSize |
Scalar or vector step size factor for finite differences. When you set
sign′(x) = sign(x) exceptsign′(0) = 1 . Central finite differences are
FiniteDifferenceStepSize expands to a vector. The default issqrt(eps) for forward finite differences, andeps^(1/3) for central finite differences.For |
FiniteDifferenceType |
Finite differences, used to estimate gradients, are either The algorithm is careful to obey bounds when estimating both types of finite differences. So, for example, it could take a backward, rather than a forward, difference to avoid evaluating at a point outside bounds. For |
|
Termination tolerance on the function value, a positive scalar. The default is For |
FunValCheck | Check whether function values are valid. |
|
Maximum number of function evaluations allowed, a positive integer. The default is For |
|
Maximum number of iterations allowed, a positive integer. The default is For |
OptimalityTolerance |
Termination tolerance on the first-order optimality (a positive scalar). The default is Internally, the For |
OutputFcn |
Specify one or more user-defined functions that an optimization function calls at each iteration. Pass a function handle or a cell array of function handles. The default is none ( |
|
情节各种措施的进展而寒冷ithm executes; select from predefined plots or write your own. Pass a name, a function handle, or a cell array of names or function handles. For custom plot functions, pass function handles. The default is none (
Custom plot functions use the same syntax as output functions. SeeOutput Functions for Optimization ToolboxandOutput Function and Plot Function Syntax. For |
SpecifyObjectiveGradient |
If For |
|
Termination tolerance on For |
|
Typical |
UseParallel |
When |
Trust-Region-Reflective Algorithm | |
JacobianMultiplyFcn |
Jacobian multiply function, specified as a function handle. For large-scale structured problems, this function computes the Jacobian matrix product W = jmfun(Jinfo,Y,flag) where [F,Jinfo] = fun(x)
In each case, Note
SeeMinimization with Dense Structured Hessian, Linear EqualitiesandJacobian Multiply Function with Linear Least Squares类似的examples. For |
JacobPattern | Sparsity pattern of the Jacobian for finite differencing. Set Use If the structure is unknown, do not set |
MaxPCGIter | Maximum number of PCG (preconditioned conjugate gradient) iterations, a positive scalar. The default is |
PrecondBandWidth | Upper bandwidth of preconditioner for PCG, a nonnegative integer. The default |
SubproblemAlgorithm |
Determines how the iteration step is calculated. The default, |
TolPCG | Termination tolerance on the PCG iteration, a positive scalar. The default is |
Levenberg-Marquardt Algorithm | |
InitDamping | Initial value of the Levenberg-Marquardt parameter, a positive scalar. Default is |
ScaleProblem |
|
Interior-Point Algorithm | |
BarrierParamUpdate |
Specifies how
This option can affect the speed and convergence of the solver, but the effect is not easy to predict. |
ConstraintTolerance |
Tolerance on the constraint violation, a positive scalar. The default is For |
InitBarrierParam | Initial barrier value, a positive scalar. Sometimes it might help to try a value above the default |
SpecifyConstraintGradient |
Gradient for nonlinear constraint functions defined by the user. When set to the default, For |
SubproblemAlgorithm |
Determines how the iteration step is calculated. The default, For |
Example:options = optimoptions('lsqnonlin','FiniteDifferenceType','central')
problem
—Problem structure
structure
Problem structure, specified as a structure with the following fields:
Field Name | Entry |
---|---|
|
Objective function |
|
Initial point forx |
|
Matrix for linear inequality constraints |
|
Vector for linear inequality constraints |
|
矩阵线性等式约束 |
|
Vector for linear equality constraints |
lb |
Vector of lower bounds |
ub |
Vector of upper bounds |
|
Nonlinear constraint function |
|
'lsqnonlin' |
|
Options created withoptimoptions |
You must supply at least theobjective
,x0
,solver
, andoptions
fields in theproblem
structure.
Data Types:struct
Output Arguments
x
— Solution
real vector | real array
Solution, returned as a real vector or real array. The size ofx
is the same as the size ofx0
. Typically,x
is a local solution to the problem whenexitflag
is positive. For information on the quality of the solution, seeWhen the Solver Succeeds.
resnorm
— Squared norm of the residual
nonnegative real
Squared norm of the residual, returned as a nonnegative real.resnorm
is the squared 2-norm of the residual atx
:sum(fun(x).^2)
.
residual
— Value of objective function at solution
array
Value of objective function at solution, returned as an array. In general,residual = fun(x)
.
exitflag
— Reason the solver stopped
integer
Reason the solver stopped, returned as an integer.
|
Function converged to a solution |
|
Change in |
|
Change in the residual is less than the specified tolerance. |
|
Relative magnitude of search direction is smaller than the step tolerance. |
|
Number of iterations exceeds |
|
A plot function or output function stopped the solver. |
|
No feasible point found. The bounds |
output
— Information about the optimization process
structure
Information about the optimization process, returned as a structure with fields:
firstorderopt |
Measure of first-order optimality |
iterations |
Number of iterations taken |
funcCount |
The number of function evaluations |
cgiterations |
Total number of PCG iterations ( |
stepsize |
Final displacement in |
constrviolation |
Maximum of constraint functions ( |
bestfeasible |
Best (lowest objective function) feasible point encountered (
If no feasible point is found, the The |
algorithm |
Optimization algorithm used |
message |
Exit message |
lambda
— Lagrange multipliers at the solution
structure
jacobian
— Jacobian at the solution
real matrix
Jacobian at the solution, returned as a real matrix.jacobian(i,j)
is the partial derivative offun(i)
with respect tox(j)
at the solutionx
.
For problems with active constraints at the solution,jacobian
is not useful for estimating confidence intervals.
Limitations
The trust-region-reflective algorithm does not solve underdetermined systems; it requires that the number of equations, i.e., the row dimension ofF, be at least as great as the number of variables. In the underdetermined case,
lsqnonlin
uses the Levenberg-Marquardt algorithm.lsqnonlin
can solve complex-valued problems directly. Note that constraints do not make sense for complex values, because complex numbers are not well-ordered; asking whether one complex value is greater or less than another complex value is nonsensical. For a complex problem with bound constraints, split the variables into real and imaginary parts. Do not use the'interior-point'
algorithm with complex data. SeeFit a Model to Complex-Valued Data.The preconditioner computation used in the preconditioned conjugate gradient part of the trust-region-reflective method formsJTJ(whereJis the Jacobian matrix) before computing the preconditioner. Therefore, a row ofJwith many nonzeros, which results in a nearly dense productJTJ, can lead to a costly solution process for large problems.
If components ofxhave no upper (or lower) bounds,
lsqnonlin
prefers that the corresponding components ofub
(orlb
) be set toinf
(or-inf
for lower bounds) as opposed to an arbitrary but very large positive (or negative for lower bounds) number.
You can use the trust-region reflective algorithm inlsqnonlin
,lsqcurvefit
, andfsolve
with small- to medium-scale problems without computing the Jacobian infun
or providing the Jacobian sparsity pattern. (This also applies to usingfmincon
orfminunc
without computing the Hessian or supplying the Hessian sparsity pattern.) How small is small- to medium-scale? No absolute answer is available, as it depends on the amount of virtual memory in your computer system configuration.
Suppose your problem hasm
equations andn
unknowns. If the commandJ = sparse(ones(m,n))
causes anOut of memory
error on your machine, then this is certainly too large a problem. If it does not result in an error, the problem might still be too large. You can find out only by running it and seeing if MATLAB runs within the amount of virtual memory available on your system.
Algorithms
The Levenberg-Marquardt and trust-region-reflective methods are based on the nonlinear least-squares algorithms also used infsolve
.
The default trust-region-reflective algorithm is a subspace trust-region method and is based on the interior-reflective Newton method described in[1]and[2]. Each iteration involves the approximate solution of a large linear system using the method of preconditioned conjugate gradients (PCG). SeeTrust-Region-Reflective Least Squares.
Levenberg-Marquardt方法中描述的裁判erences[4],[5], and[6]. SeeLevenberg-Marquardt Method.
The'interior-point'
algorithm uses thefmincon
'interior-point'
algorithm with some modifications. For details, seeModified fmincon Algorithm for Constrained Least Squares.
Alternative Functionality
App
TheOptimizeLive Editor task provides a visual interface forlsqnonlin
.
References
[1] Coleman, T.F. and Y. Li. “An Interior, Trust Region Approach for Nonlinear Minimization Subject to Bounds.”SIAM Journal on Optimization, Vol. 6, 1996, pp. 418–445.
[2] Coleman, T.F. and Y. Li. “On the Convergence of Reflective Newton Methods for Large-Scale Nonlinear Minimization Subject to Bounds.”Mathematical Programming, Vol. 67, Number 2, 1994, pp. 189–224.
[3] Dennis, J. E. Jr. “Nonlinear Least-Squares.”State of the Art in Numerical Analysis, ed. D. Jacobs, Academic Press, pp. 269–312.
[4] Levenberg, K. “A Method for the Solution of Certain Problems in Least-Squares.”Quarterly Applied Mathematics 2, 1944, pp. 164–168.
[5] Marquardt, D. “An Algorithm for Least-squares Estimation of Nonlinear Parameters.”SIAM Journal Applied Mathematics, Vol. 11, 1963, pp. 431–441.
[6] Moré, J. J. “The Levenberg-Marquardt Algorithm: Implementation and Theory.”Numerical Analysis, ed. G. A. Watson, Lecture Notes in Mathematics 630, Springer Verlag, 1977, pp. 105–116.
[7] Moré, J. J., B. S. Garbow, and K. E. Hillstrom.User Guide for MINPACK 1. Argonne National Laboratory, Rept. ANL–80–74, 1980.
[8] Powell, M. J. D. “A Fortran Subroutine for Solving Systems of Nonlinear Algebraic Equations.”Numerical Methods for Nonlinear Algebraic Equations, P. Rabinowitz, ed., Ch.7, 1970.
Extended Capabilities
C/C++ Code Generation
Generate C and C++ code using MATLAB® Coder™.
lsqcurvefit
andlsqnonlin
support code generation using either thecodegen
(MATLAB Coder)function or theMATLAB编码器™app. You must have aMATLAB编码器license to generate code.The target hardware must support standard double-precision floating-point computations. You cannot generate code for single-precision or fixed-point computations.
Code generation targets do not use the same math kernel libraries as MATLAB solvers. Therefore, code generation solutions can vary from solver solutions, especially for poorly conditioned problems.
All code for generation must be MATLAB code. In particular, you cannot use a custom black-box function as an objective function for
lsqcurvefit
orlsqnonlin
. You can usecoder.ceval
to evaluate a custom function coded in C or C++. However, the custom function must be called in a MATLAB function.Code generation for
lsqcurvefit
andlsqnonlin
currently does not support linear or nonlinear constraints.lsqcurvefit
andlsqnonlin
do not support theproblem
论点for code generation.[x,fval] = lsqnonlin(problem)% Not supported
You must specify the objective function by using function handles, not strings or character names.
x = lsqnonlin(@fun,x0,lb,ub,options)% Supported% Not supported: lsqnonlin('fun',...) or lsqnonlin("fun",...)
All input matrices
lb
andub
must be full, not sparse. You can convert sparse matrices to full by using thefull
function.The
lb
andub
论点s must have the same number of entries as thex0
论点or must be empty[]
.If your target hardware does not support infinite bounds, use
optim.coder.infbound
.For advanced code optimization involving embedded processors, you also need an Embedded Coder®license.
You must include options for
lsqcurvefit
orlsqnonlin
and specify them usingoptimoptions
. The options must include theAlgorithm
option, set to'levenberg-marquardt'
.options = optimoptions('lsqnonlin','Algorithm','levenberg-marquardt'); [x,fval,exitflag] = lsqnonlin(fun,x0,lb,ub,options);
Code generation supports these options:
Algorithm
— Must be'levenberg-marquardt'
FiniteDifferenceStepSize
FiniteDifferenceType
FunctionTolerance
MaxFunctionEvaluations
MaxIterations
SpecifyObjectiveGradient
StepTolerance
TypicalX
Generated code has limited error checking for options. The recommended way to update an option is to use
optimoptions
, not dot notation.opts = optimoptions('lsqnonlin','Algorithm','levenberg-marquardt'); opts = optimoptions(opts,'MaxIterations',1e4);% Recommendedopts.MaxIterations = 1e4;% Not recommended
Do not load options from a file. Doing so can cause code generation to fail. Instead, create options in your code.
Usually, if you specify an option that is not supported, the option is silently ignored during code generation. However, if you specify a plot function or output function by using dot notation, code generation can issue an error. For reliability, specify only supported options.
Because output functions and plot functions are not supported, solvers do not return the exit flag –1.
For an example, seeGenerate Code for lsqcurvefit or lsqnonlin.
Automatic Parallel Support
Accelerate code by automatically running computation in parallel using Parallel Computing Toolbox™.
To run in parallel, set the'UseParallel'
option totrue
.
options = optimoptions('
solvername
','UseParallel',true)
For more information, seeUsing Parallel Computing in Optimization Toolbox.
Version History
Introduced before R2006aR2023a:Linear and Nonlinear Constraint Support
lsqnonlin
gains support for both linear and nonlinear constraints. To enable constraint satisfaction, the solver uses the"interior-point"
algorithm fromfmincon
.
If you specify constraints but do not specify an algorithm, the solver automatically switches to the
"interior-point"
algorithm.If you specify constraints and an algorithm, you must specify the
"interior-point"
algorithm.
For algorithm details, seeModified fmincon Algorithm for Constrained Least Squares. For an example, see比较lsqnonlin和fmincon Nonl受限inear Least Squares.
Open Example
You have a modified version of this example. Do you want to open this example with your edits?
MATLAB Command
You clicked a link that corresponds to this MATLAB command:
Run the command by entering it in the MATLAB Command Window. Web browsers do not support MATLAB commands.
Select a Web Site
Choose a web site to get translated content where available and see local events and offers. Based on your location, we recommend that you select:.
You can also select a web site from the following list:
How to Get Best Site Performance
Select the China site (in Chinese or English) for best site performance. Other MathWorks country sites are not optimized for visits from your location.
Americas
- América Latina(Español)
- Canada(English)
- United States(English)
Europe
- Belgium(English)
- Denmark(English)
- Deutschland(Deutsch)
- España(Español)
- Finland(English)
- France(Français)
- Ireland(English)
- Italia(Italiano)
- Luxembourg(English)
- Netherlands(English)
- Norway(English)
- Österreich(Deutsch)
- Portugal(English)
- Sweden(English)
- Switzerland
- United Kingdom(English)