I have a histogram that seems to fit a poisson distribution.
In order to fit it, I declare the function myself as follows
xdata; ydata; % Arrays in which I have stored the data.
%Ydata tell us how many times the xdata is repeated in the set.
fun= #(x,xdata) (exp(-x(1))*(x(1).^(xdata)) )/(factorial(xdata)) %Function I
% want to use in the fit. It is a poisson distribution.
x0=[1]; %Approximated value of the parameter lambda to help the fit
p=lsqcurvefit(fun,x0,xdata,ydata); % Fit in the least square sense
I find an error. It probably has to do with the "factorial". Any ideas?
Factorial outputs a vector from vector xdata. Why are you using .xdata in factorial?
For example:
data = [1 2 3];
factorial(data) is then [1! 2! 3!].
Try ./factorial(xdata) (I cannot recall if the dot is even necessary at this case.)
You need to use gamma(xdata+1) function instead of factorial(xdata) function. Gamma function is a generalized form of factorial function which can be used for real and complex numbers. Thus, your code would be:
fun = #(x,xdata) exp(-x(1))*x(1).^xdata./gamma(xdata+1);
x = lsqcurvefit(fun,1,xdata,ydata);
Alternatively, you can MATLAB fitdist function which is already optimized and you might get better results:
pd = fitdist(xdata,'Poisson','Frequency',ydata);
pd.lambda
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.
Let's say I want to model this equation (electrical motor, 6 phases):
Vs = Rs*Is + d/dt*(Ls*Is)
where all variables are matrix, so:
Vs = [va1 vb1 vc1 va2 vb2 vc2]' (column vector)
Is = [ia1 ib1 ic1 ia2 ib2 ic2]' (column vector)
Ls and Rs are 6x6 matrix (constants)
From my point of view the Vs is the input vector and Is is the output vector so I need to rearrange the equation.
I have seen that is not possible in Simulink to feed the Transfer Fcn block with matrix, at least not for a multiple input multiple output system.
Is there a way to realize this on Simulink still using the matrix Ls and Rs without "unpacking" the equation?
Thank you
I would re-arrange your equations in state-space form and use the State-Space block, which is better suited for matrix equations.
Another option is to use basic Simulink blocks, such as Integrator and Gain blocks, with vectorized inputs. I am not 100% sure this will work, but reasonably confident.
You can use the product block with matrices inside Matlab so there should be no problem. It's also possible to use integrator/derivative block (though it's better to avoid using derivative if possible) with a vector input so if you can put your equation with Is as an output there should be no problem.
You can put 2 multiply blocks with your matrices as inputs and the vector you need and you will get a vector for the output like you want.
My Matlab program has multiple inputs as a struct (in.a, in.b, etc.)
and multiple outputs (out.a, out.b, etc.)
I would like to use the genetic algorithm solver from teh optimization toolbox to find the best input in.a, while all the other inputs are constant. The fitness is one of the outputs, e.g. out.b(2,3).
How do I "tell" the solver this?
Thanks
Daniel
It is not uncommon in programming to have a situation where what is most convenient for your function and what some library call expects of it don't agree. The normal resolution to such a problem is to write a small layer in between that allows the two to talk; an interface.
From help ga:
X = GA(FITNESSFCN,NVARS) finds a local unconstrained minimum X to the
FITNESSFCN using GA. [...] FITNESSFCN accepts a vector X of size
1-by-NVARS, and returns a scalar evaluated at X.
So, ga expects vector input, scalar output, whereas you have a structure going in and out. You would have to write the following (sub)function:
function Y = wrapper_Objfun(X, in)
in.a = X; %# variable being optimized
out = YOUR_REAL_FUNCTION(in); %# call to your actual function
Y = out.b(2,3); %# objective value
end
and then the call to ga will look like
X = ga(#(x) wrapper_Objfun(x,in), N);
where N is however large in.a should be.
Also have a read about it in Matlab's own documentation on the subject.
If I have an RC circuit with transfer function 1/(1+sRC) how do I draw the transfer function using MATLAB?
Num2=[1];
Den2=[R*C 1];
RCcirc=tf(Num2,Den2);
How do I declare the R and the C so that there are no errors?
tf is the wrong tool for plotting the transfer function. Try these instead:
Use linspace to generate a range of values for s. Give R and C reasonable values of your choice.
Read up on arithmetic operations in MATLAB, especially ./
Look at how to use plot and familiarize yourself with the command using some simple examples from the docs.
With these you should be able to plot the transfer function in MATLAB :)
First of all you need to understand what transfer function you want. Without defined values of R and C you won't get any transfer function. Compare it to this, you want to plot a sine wave: x = sin(w*t), I hope you can agree with me that you cannot plot such a function (including axes) unless I specifically say e.g. t is the time, ranging from 0 seconds to 10 seconds and w is a pulsation of 1 rad/s. It's exactly the same with your RC network: without any values, it is impossible for numerical software such as MATLAB to come up with a plot.
If you fill in those values, you can use th tf function to display the transfer function in whatever way you like (e.g. a bode plot).
On the other hand, if you just want the expression 1/(1+s*R*C), take a look at the symbolic toolbox, you can do such things there. But to make a plot, you will still have to fill in the R and C value (and even a value for your Laplace variable in this case).