I'm using MATLAB's fsolve function to solve systems of nonlinear equations. I have two nonlinear equations with two variables (x,y);
I'm trying to find the all possible roots for the both variables. I noted that the fsolve gives just one root. How it possible to get the all roots for the equations?
My code as the following:
function F = fun(guess)
x = guess(1);
y = guess(2);
F = [2*x -y - exp(-x));
-x + 2*y - exp(-y) ];
end
call the function:
guess = [-5 -5]
fsolve(#fun,guess);
Prove that there is only one root, so you don't need to search further.
From the second equation,
-x + 2·y - exp(-y) = 0
⇒ x = 2·y - exp(-y)
Substitute x into the first equation:
2·x - y - exp(-x) = 0
⇒ 2·(2y-exp(-y)) - y - exp(-(2y-exp(-y)) = 0
which is a function of y only. Standard calculus will show this f(y) is monotonically increasing, starts negative at y=-∞ and is positive at y=+∞. The result is the same for when you do the substitution the other way around. This implies there is only 1 simultaneous root to both equations.
QED.
fsolve is not a global solver. There are global solvers (like genetic algorithms and simulated annealing), but they have to run for an infinite amount of time to guarantee that the returned solutions comprise all of the minimizers. On the hand, almost all other optimization solvers are local, meaning that they will only guarantee that a local minimizer is returned.
Furthermore, in addition to not knowing whether the returned solution is a global or local minimizer, there is, in general, no way of determining how many roots a problem has. So basically, there is no way of doing what you want except for 2 well known cases:
1) If the problem is convex, then there are no local minimizers that are not global minimizers. So anything returned by fsolve will be a global minimizer. Furthermore, this minimizer is almost always unique. The exception is that technically, there can exist an infinite number of solutions, but they will all be connected (like on a specific plane). There cannot exist a finite number of distinct minimizers that are not connected.
2) Polynomials have a distinct number of roots that we can uniquely determine.
Related
I'm solving a pair of non-linear equations for each voxel in a dataset of a ~billion voxels using fsolve() in MATLAB 2016b.
I have done all the 'easy' optimizations that I'm aware of. Memory localization is OK, I'm using parfor, the equations are in fairly numerically simple form. All discontinuities of the integral are fed to integral(). I'm using the Levenberg-Marquardt algorithm with good starting values and a suitable starting damping constant, it converges on average with 6 iterations.
I'm now at ~6ms per voxel, which is good, but not good enough. I'd need a order of magnitude reduction to make the technique viable. There's only a few things that I can think of improving before starting to hammer away at accuracy:
The splines in the equation are for quick sampling of complex equations. There are two for each equation, one is inside the 'complicated nonlinear equation'. They represent two equations, one which is has a large amount of terms but is smooth and has no discontinuities and one which approximates a histogram drawn from a spectrum. I'm using griddedInterpolant() as the editor suggested.
Is there a faster way to sample points from pre-calculated distributions?
parfor i=1:numel(I1)
sols = fsolve(#(x) equationPair(x, input1, input2, ...
6 static inputs, fsolve options)
output1(i) = sols(1); output2(i) = sols(2)
end
When calling fsolve, I'm using the 'parametrization' suggested by Mathworks to input the variables. I have a nagging feeling that defining a anonymous function for each voxel is taking a large slice of the time at this point. Is this true, is there a relatively large overhead for defining the anonymous function again and again? Do I have any way to vectorize the call to fsolve?
There are two input variables which keep changing, all of the other input variables stay static. I need to solve one equation pair for each input pair so I can't make it a huge system and solve it at once. Do I have any other options than fsolve for solving pairs of nonlinear equations?
If not, some of the static inputs are the fairly large. Is there a way to keep the inputs as persistent variables using MATLAB's persistent, would that improve performance? I only saw examples of how to load persistent variables, how could I make it so that they would be input only once and future function calls would be spared from the assumedly largish overhead of the large inputs?
EDIT:
The original equations in full form look like:
Where:
and:
Everything else is known, solving for x_1 and x_2. f_KN was approximated by a spline. S_low (E) and S_high(E) are splines, the histograms they are from look like:
So, there's a few things I thought of:
Lookup table
Because the integrals in your function do not depend on any of the parameters other than x, you could make a simple 2D-lookup table from them:
% assuming simple (square) range here, adjust as needed
[x1,x2] = meshgrid( linspace(0, xmax, N) );
LUT_high = zeros(size(x1));
LUT_low = zeros(size(x1));
for ii = 1:N
LUT_high(:,ii) = integral(#(E) Fhi(E, x1(1,ii), x2(:,ii)), ...
0, E_high, ...
'ArrayValued', true);
LUT_low(:,ii) = integral(#(E) Flo(E, x1(1,ii), x2(:,ii)), ...
0, E_low, ...
'ArrayValued', true);
end
where Fhi and Flo are helper functions to compute those integrals, vectorized with scalar x1 and vector x2 in this example. Set N as high as memory will allow.
Those lookup tables you then pass as parameters to equationPair() (which allows parfor to distribute the data). Then just use interp2 in equationPair():
F(1) = I_high - interp2(x1,x2,LUT_high, x(1), x(2));
F(2) = I_low - interp2(x1,x2,LUT_low , x(1), x(2));
So, instead of recomputing the whole integral every time, you evaluate it once for the expected range of x, and reuse the outcomes.
You can specify the interpolation method used, which is linear by default. Specify cubic if you're really concerned about accuracy.
Coarse/Fine
Should the lookup table method not be possible for some reason (memory limitations, in case the possible range of x is too big), here's another thing you could do: split up the whole procedure in 2 parts, which I'll call coarse and fine.
The intent of the coarse method is to improve your initial estimates really quickly, but perhaps not so accurately. The quickest way to approximate that integral by far is via the rectangle method:
do not approximate S with a spline, just use the original tabulated data (so S_high/low = [S_high/low#E0, S_high/low#E1, ..., S_high/low#E_high/low]
At the same values for E as used by the S data (E0, E1, ...), evaluate the exponential at x:
Elo = linspace(0, E_low, numel(S_low)).';
integrand_exp_low = exp(x(1)./Elo.^3 + x(2)*fKN(Elo));
Ehi = linspace(0, E_high, numel(S_high)).';
integrand_exp_high = exp(x(1)./Ehi.^3 + x(2)*fKN(Ehi));
then use the rectangle method:
F(1) = I_low - (S_low * Elo) * (Elo(2) - Elo(1));
F(2) = I_high - (S_high * Ehi) * (Ehi(2) - Ehi(1));
Running fsolve like this for all I_low and I_high will then have improved your initial estimates x0 probably to a point close to "actual" convergence.
Alternatively, instead of the rectangle method, you use trapz (trapezoidal method). A tad slower, but possibly a bit more accurate.
Note that if (Elo(2) - Elo(1)) == (Ehi(2) - Ehi(1)) (step sizes are equal), you can further reduce the number of computations. In that case, the first N_low elements of the two integrands are identical, so the values of the exponentials will only differ in the N_low + 1 : N_high elements. So then just compute integrand_exp_high, and set integrand_exp_low equal to the first N_low elements of integrand_exp_high.
The fine method then uses your original implementation (with the actual integral()s), but then starting at the updated initial estimates from the coarse step.
The whole objective here is to try and bring the total number of iterations needed down from about 6 to less than 2. Perhaps you'll even find that the trapz method already provides enough accuracy, rendering the whole fine step unnecessary.
Vectorization
The rectangle method in the coarse step outlined above is easy to vectorize:
% (uses R2016b implicit expansion rules)
Elo = linspace(0, E_low, numel(S_low));
integrand_exp_low = exp(x(:,1)./Elo.^3 + x(:,2).*fKN(Elo));
Ehi = linspace(0, E_high, numel(S_high));
integrand_exp_high = exp(x(:,1)./Ehi.^3 + x(:,2).*fKN(Ehi));
F = [I_high_vector - (S_high * integrand_exp_high) * (Ehi(2) - Ehi(1))
I_low_vector - (S_low * integrand_exp_low ) * (Elo(2) - Elo(1))];
trapz also works on matrices; it will integrate over each column in the matrix.
You'd call equationPair() then using x0 = [x01; x02; ...; x0N], and fsolve will then converge to [x1; x2; ...; xN], where N is the number of voxels, and each x0 is 1×2 ([x(1) x(2)]), so x0 is N×2.
parfor should be able to slice all of this fairly easily over all the workers in your pool.
Similarly, vectorization of the fine method should also be possible; just use the 'ArrayValued' option to integral() as shown above:
F = [I_high_vector - integral(#(E) S_high(E) .* exp(x(:,1)./E.^3 + x(:,2).*fKN(E)),...
0, E_high,...
'ArrayValued', true);
I_low_vector - integral(#(E) S_low(E) .* exp(x(:,1)./E.^3 + x(:,2).*fKN(E)),...
0, E_low,...
'ArrayValued', true);
];
Jacobian
Taking derivatives of your function is quite easy. Here is the derivative w.r.t. x_1, and here w.r.t. x_2. Your Jacobian will then have to be a 2×2 matrix
J = [dF(1)/dx(1) dF(1)/dx(2)
dF(2)/dx(1) dF(2)/dx(2)];
Don't forget the leading minus sign (F = I_hi/lo - g(x) → dF/dx = -dg/dx)
Using one or both of the methods outlined above, you can implement a function to compute the Jacobian matrix and pass this on to fsolve via the 'SpecifyObjectiveGradient' option (via optimoptions). The 'CheckGradients' option will come in handy there.
Because fsolve usually spends the vast majority of its time computing the Jacobian via finite differences, manually computing a value for it manually will normally speed the algorithm up tremendously.
It will be faster, because
fsolve doesn't have to do extra function evaluations to do the finite differences
the convergence rate will increase due to the improved precision of the Jacobian
Especially if you use the rectangle method or trapz like above, you can reuse many of the computations you've already done for the function values themselves, meaning, even more speed-up.
Rody's answer was the correct one. Supplying the Jacobian was the single largest factor. Especially with the vectorized version, there were 3 orders of magnitude of difference in speed with the Jacobian supplied and not.
I had trouble finding information about this subject online so I'll spell it out here for future reference: It is possible to vectorize independant parallel equations with fsolve() with great gains.
I also did some work with inlining fsolve(). After supplying the Jacobian and being smarter about the equations, the serial version of my code was mostly overhead at ~1*10^-3 s per voxel. At that point most of the time inside the function was spent passing around a options -struct and creating error-messages which are never sent + lots of unused stuff assumedly for the other optimization functions inside the optimisation function (levenberg-marquardt for me). I succesfully butchered the function fsolve and some of the functions it calls, dropping the time to ~1*10^-4s per voxel on my machine. So if you are stuck with a serial implementation e.g. because of having to rely on the previous results it's quite possible to inline fsolve() with good results.
The vectorized version provided the best results in my case, with ~5*10^-5 s per voxel.
I have a system of equations...
dF(a,b,c)/da = 0;
dF(a,b,c)/db = 0;
dF(a,b,c)/dc = 0;
where a,b,c are unknown variable constants and dF/d* are anonymous functions of the variables. I have to solve for a,b and c in an optimization problem. When the system reduces to just one equation, I use Matlab's fzero to solve for the variable and it works. For example
var_a = fzero(#(a) dF(a)/da,0);
After noticing that fzero and fsolve give dramatically different answers for some cases I did some searching. From what I gather, fzero only works for a single equation of a single variable? So moving to a system of equations, I'd like to choose the most appropriate method. I've used Matlab's solve in the past, but I believe that is for symbolic expressions only? What is the best method for solving a linear system of anonymous functions, which all equal zero?
I tried the following, and got back results
vars = fsolve(#(V)[dF(V)/da;dF(V)/db;dF(V)/dc],zeros(1,3));
where vars contains all 3 variables, but after reading the examples in the previous link, Fsolve couldn't exactly find the zeros for x^2 and x^3. The solution vector in the system I presented above is all zeros and the functions are polynomials. Putting this all together, I'm wondering if fsolve isn't the best choice?
Can I build a system of calls to fzero? Something along the lines of
vars = [fzero(#(a) dF(a,b,c)/da,0);
fzero(#(b) dF(a,b,c)/db,0);
fzero(#(c) dF(a,b,c)/dc,0)];
which I don't think would work (how would each dF/d* get the other 2 variable inputs?) or would it?
Any thoughts?
You can numerically solve to minimize any function using 'lsqnonlin'. To adopt this for a system of equations, simply turn them into a single function with a vector input. Something like this:
fToMinimize = #(abc) ...
(dF(ABC(1),ABC(2),ABC(3))/da)^2 +...
(dF(ABC(1),ABC(2),ABC(3))/db)^2 +...
(dF(ABC(1),ABC(2),ABC(3))/dc)^2 ;
abcSolved = lsqnonlin(fToMinimize, [0 0 0])
If you have a guess for the values of a, b, and c, you can (and should) use those instead of the [0 0 0] vector. There are also many options within the lsqnonlin function to adjust behavior. For example how close to the best answer you want to get. If the functions are well behaved, you should be able to tighten the tolerance down a lot, if you are looking for a near exact answer.
I have an integration function which does not have indefinite integral expression.
Specifically, the function is f(y)=h(y)+integral(#(x) exp(-x-1/x),0,y) where h(y) is a simple function.
Matlab numerically computes f(y) well, but I want to compute the following function.
g(w)=w*integral(1-f(y).^(1/w),0,inf) where w is a real number in [0,1].
The problem for computing g(w) is handling f(y).^(1/w) numerically.
How can I calculate g(w) with MATLAB? Is it impossible?
Expressions containing e^(-1/x) are generally difficult to compute near x = 0. Actually, I am surprised that Matlab computes f(y) well in the first place. I'd suggest trying to compute g(w)=w*integral(1-f(y).^(1/w),epsilon,inf) for epsilon greater than zero, then gradually decreasing epsilon toward 0 to check if you can get numerical convergence at all. Convergence is certainly not guaranteed!
You can calculate g(w) using the functions you have, but you need to add the (ArrayValued,true) name-value pair.
The option allows you to specify a vector-valued w and allows the nested integral call to receive a vector of y values, which is how integral naturally works.
f = #(y) h(y)+integral(#(x) exp(-x-1/x),0,y,'ArrayValued',true);
g = #(w) w .* integral(1-f(y).^(1./w),0,Inf,'ArrayValued',true);
At least, that works on my R2014b installation.
Note: While h(y) may be simple, if it's integral over the positive real line does not converge, g(w) will more than likely not converge (I don't think I need to qualify that, but I'll hedge my bets).
I have been working on solving some equation in a more complicated context. However, I want to illustrate my question through the following simple example.
Consider the following two functions:
function y=f1(x)
y=1-x;
end
function y=f2(x)
if x<0
y=0;
else
y=x;
end
end
I want to solve the following equation: f1(x)=f2(x). The code I used is:
syms x;
x=solve(f1(x)-f2(x));
And I got the following error:
??? Error using ==> sym.sym>notimplemented at 2621
Function 'lt' is not implemented for MuPAD symbolic objects.
Error in ==> sym.sym>sym.lt at 812
notimplemented('lt');
Error in ==> f2 at 3
if x<0
I know the error is because x is a symbolic variable and therefore I could not compare x with 0 in the piecewise function f2(x).
Is there a way to fix this and solve the equation?
First, make sure symbolic math is even the appropriate solution method for your problem. In many cases it isn't. Look at fzero and fsolve amongst many others. A symbolic method is only needed if, for example, you want a formula or if you need to ensure precision.
In such an old version of Matlab, you may want to break up your piecewise function into separate continuous functions and solve them separately:
syms x;
s1 = solve(1-x^2,x) % For x >= 0
s2 = solve(1-x,x) % For x < 0
Then you can either manually examine or numerically compare the outputs to determine if any or all of the solutions are valid for the chosen regime – something like this:
s = [s1(double(s1) >= 0);s2(double(s2) < 0)]
You can also take advantage of the heaviside function, which is available in much older versions.
syms x;
f1 = 1-x;
f2 = x*heaviside(x);
s = solve(f1-f2,x)
Yes, the Heaviside function is 0.5 at zero – this gives it the appropriate mathematical properties. You can shift it to compare values other than zero. This is a standard technique.
In Matlab R2012a+, you can take advantage of assumptions in addition to the normal relational operators. To add to #AlexB's comment, you should convert the output of any logical comparison to symbolic before using isAlways:
isAlways(sym(x<0))
In your case, x is obviously not "always" on one side or the other of zero, but you may still find this useful in other cases.
If you want to get deep into Matlab's symbolic math, you can create piecewise functions using MuPAD, which are accessible from Matlab – e.g., see my example here.
I'm trying to do some parameter estimation and want to choose parameter estimates that minimize the square error in a predicted equation over about 30 variables. If the equation were linear, I would just compute the 30 partial derivatives, set them all to zero, and use a linear-equation solver. But unfortunately the equation is nonlinear and so are its derivatives.
If the equation were over a single variable, I would just use Newton's method (also known as Newton-Raphson). The Web is rich in examples and code to implement Newton's method for functions of a single variable.
Given that I have about 30 variables, how can I program a numeric solution to this problem using Newton's method? I have the equation in closed form and can compute the first and second derivatives, but I don't know quite how to proceed from there. I have found a large number of treatments on the web, but they quickly get into heavy matrix notation. I've found something moderately helpful on Wikipedia, but I'm having trouble translating it into code.
Where I'm worried about breaking down is in the matrix algebra and matrix inversions. I can invert a matrix with a linear-equation solver but I'm worried about getting the right rows and columns, avoiding transposition errors, and so on.
To be quite concrete:
I want to work with tables mapping variables to their values. I can write a function of such a table that returns the square error given such a table as argument. I can also create functions that return a partial derivative with respect to any given variable.
I have a reasonable starting estimate for the values in the table, so I'm not worried about convergence.
I'm not sure how to write the loop that uses an estimate (table of value for each variable), the function, and a table of partial-derivative functions to produce a new estimate.
That last is what I'd like help with. Any direct help or pointers to good sources will be warmly appreciated.
Edit: Since I have the first and second derivatives in closed form, I would like to take advantage of them and avoid more slowly converging methods like simplex searches.
The Numerical Recipes link was most helpful. I wound up symbolically differentiating my error estimate to produce 30 partial derivatives, then used Newton's method to set them all to zero. Here are the highlights of the code:
__doc.findzero = [[function(functions, partials, point, [epsilon, steps]) returns table, boolean
Where
point is a table mapping variable names to real numbers
(a point in N-dimensional space)
functions is a list of functions, each of which takes a table like
point as an argument
partials is a list of tables; partials[i].x is the partial derivative
of functions[i] with respect to 'x'
epilson is a number that says how close to zero we're trying to get
steps is max number of steps to take (defaults to infinity)
result is a table like 'point', boolean that says 'converged'
]]
-- See Numerical Recipes in C, Section 9.6 [http://www.nrbook.com/a/bookcpdf.php]
function findzero(functions, partials, point, epsilon, steps)
epsilon = epsilon or 1.0e-6
steps = steps or 1/0
assert(#functions > 0)
assert(table.numpairs(partials[1]) == #functions,
'number of functions not equal to number of variables')
local equations = { }
repeat
if Linf(functions, point) <= epsilon then
return point, true
end
for i = 1, #functions do
local F = functions[i](point)
local zero = F
for x, partial in pairs(partials[i]) do
zero = zero + lineq.var(x) * partial(point)
end
equations[i] = lineq.eqn(zero, 0)
end
local delta = table.map(lineq.tonumber, lineq.solve(equations, {}).answers)
point = table.map(function(v, x) return v + delta[x] end, point)
steps = steps - 1
until steps <= 0
return point, false
end
function Linf(functions, point)
-- distance using L-infinity norm
assert(#functions > 0)
local max = 0
for i = 1, #functions do
local z = functions[i](point)
max = math.max(max, math.abs(z))
end
return max
end
You might be able to find what you need at the Numerical Recipes in C web page. There is a free version available online. Here (PDF) is the chapter containing the Newton-Raphson method implemented in C. You may also want to look at what is available at Netlib (LINPack, et. al.).
As an alternative to using Newton's method the Simplex Method of Nelder-Mead is ideally suited to this problem and referenced in Numerical Recpies in C.
Rob
You are asking for a function minimization algorithm. There are two main classes: local and global. Your problem is least squares so both local and global minimization algorithms should converge to the same unique solution. Local minimization is far more efficient than global so select that.
There are many local minimization algorithms but one particularly well suited to least squares problems is Levenberg-Marquardt. If you don't have such a solver to hand (e.g. from MINPACK) then you can probably get away with Newton's method:
x <- x - (hessian x)^-1 * grad x
where you compute the inverse matrix multiplied by a vector using a linear solver.
Since you already have the partial derivatives, how about a general gradient-descent approach?
Maybe you think you have a good-enough solution, but for me, the easiest way to think about this is to understand it in the 1-variable case first, and then extend it to the matrix case.
In the 1-variable case, if you divide the first derivative by the second derivative, you get the (negative) step size to your next trial point, e.g. -V/A.
In the N-variable case, the first derivative is a vector and the second derivative is a matrix (the Hessian). You multiply the derivative vector by the inverse of the second derivative, and the result is the negative step-vector to your next trial point, e.g. -V*(1/A)
I assume you can get the 2nd-derivative Hessian matrix. You will need a routine to invert it. There are plenty of these around in various linear algebra packages, and they are quite fast.
(For readers who are not familiar with this idea, suppose the two variables are x and y, and the surface is v(x,y). Then the first derivative is the vector:
V = [ dv/dx, dv/dy ]
and the second derivative is the matrix:
A = [dV/dx]
[dV/dy]
or:
A = [ d(dv/dx)/dx, d(dv/dy)/dx]
[ d(dv/dx)/dy, d(dv/dy)/dy]
or:
A = [d^2v/dx^2, d^2v/dydx]
[d^2v/dxdy, d^2v/dy^2]
which is symmetric.)
If the surface is parabolic (constant 2nd derivative) it will get to the answer in 1 step. On the other hand, if the 2nd derivative is very not-constant, you could encounter oscillation. Cutting each step in half (or some fraction) should make it stable.
If N == 1, you'll see that it does the same thing as in the 1-variable case.
Good luck.
Added: You wanted code:
double X[N];
// Set X to initial estimate
while(!done){
double V[N]; // 1st derivative "velocity" vector
double A[N*N]; // 2nd derivative "acceleration" matrix
double A1[N*N]; // inverse of A
double S[N]; // step vector
CalculateFirstDerivative(V, X);
CalculateSecondDerivative(A, X);
// A1 = 1/A
GetMatrixInverse(A, A1);
// S = V*(1/A)
VectorTimesMatrix(V, A1, S);
// if S is small enough, stop
// X -= S
VectorMinusVector(X, S, X);
}
My opinion is to use a stochastic optimizer, e.g., a Particle Swarm method.