My goal is to solve for a matrix [A] that satisfies [A]*[B]=[C] where [C] is known and [B] is randomly generated. Below is an example:
C=[1/3 1/3 1/3]'*[1/3 1/6 1/6 1/6 1/6];
B=rand(5,5);
A=C*pinv(B);
A*B=C_test;
norm(C-C_test);
ans =
4.6671e-16
Here the elements of [C_test] are within 1e-15 to the original [C], but when [B] has less rows than columns, the error dramatically increases (not sure is norm() is the best way to show the error, but I think it illustrates the problem). For example:
B=rand(4,5);
A=C*pinv(B);
A*B=C_test;
norm(C-C_test);
ans =
0.0173
Additional methods:
QR-Factorization
[Q,R,P]=qr(B);
A=((C*P)/R))*Q';
norm(C-A*B);
ans =
0.0173
/ Operator
A=C/B;
norm(C-A*B);
ans =
0.0173
Why does this happen? In both cases [B]*pinv([B])=[I] so it seems like the process should work.
If this is a numerical or algebraic fact of life associated with pinv() or the other methods, is there another way I can generate [A] to satisfy the equation? Thank you!
Since C is 3×5, the number of elements in C and hence the number of equations is equal to 15. If B is 5×5, the number of unknowns (the elements in A) equals 3×5 = 15 as well, and the solution will be accurate.
If on the other hand B is for instance 3×5, the number of elements in A is equal to 3×3 = 9 and hence the system is overdetermined, which means that the resulting A will be the least-squares solution.
See for general information wikipedia: System of linear equations, and Matlabs Overdetermined system.
The resulting matrix A is the best fit and there is no way to improve (in a least square sense).
In response to your second question: you are measuring the quality of A*B as an approximation of C by applying the 2-norm to A*B-C: which is equivalent to least-squares fitting. In this measure, all the approaches that you use provide the optimal answer.
If you however would prefer some other measure, such as the 1-norm, the Infinity-norm or any other measure (for instance by picking different weights for column, row or element), the obtained answers from the original approach will of course not be necessarily optimal with respect to this new measure.
The most general approach would be to use some optimization routine, like this:
x = fminunc(f, zeros(3*size(B,1),1));
A = reshape(x,3,size(B,1));
where f is some (any) measure. The least-square measure should result in the same A. So if you try this one:
f = #(x) norm(reshape(x,3,size(B,1))*B - C);
A should match the results in your approaches.
But you could use any f here. For instance, try the 1-norm:
f = #(x) norm(reshape(x,3,size(B,1))*B - C, 1);
Or something crazy like:
f = #(x) sum(abs(reshape(x,3,size(B,1))*B - C)*[1 10 100 1000 10000]');
This will give different results, which are according to the new measure f optimal. That being said, I would stick to the least squares ;)
Related
I've been trying to solve whether two m-by-n matrices A and B are equivalent via
[a,ja] = rref(A,tol)
[b,jb] = rref(B,tol)
and then comparing
isequal(a,b) & isequal(ja,jb)
First, I don't really understand what ja and jb are. My problem is that the reduced row echelon form is very simple for both A and B and identical in all cases. I don't know whether this is on purpose or not. For example, I get equivalence for just
A = rand(40,3)
B = rand(40,3)
which I'm not sure is nonsense or not.
It looks like you're trying to check if the reduced row echelon forms of two matrices are element-wise equivalent. From how you've defined A and B, they are (it's effectively an overdetermined system, I believe). However, I think that you may have flipped your rows and columns. If instead you create A and B such that there are more columns than rows (i.e., an augmented matrix of an underdetermined system):
A = rand(3,40)
B = rand(3,40)
then when you run rref, you'll see a much different output and your comparison will return false as you perhaps expected.
Also, I think that it is sufficient to use the following, as two matrices that are equal element-wise will surely share the same rank (or approximation thereof):
a = rref(A,tol);
b = rref(B,tol);
isequal(a,b)
Lets say I have two vectors A and B with different lengths Length(A) is not equal to Length(B) and the Values in Vector A, are not the same as in Vector B. I want to compare each value of B with Values of A (Compare means if Value B(i) is almost the same value of A(1:end) for example B(i)-Tolerance<A(i)<B(i)+Tolerance.
How Can I do this without using for loop since the data is huge?
I know ismember(F), intersect,repmat,find but non of those function can really help me
You may try a solution along these lines:
tol = 0.1;
N = 1000000;
a = randn(1, N)*1000; % create a randomly
b = a + tol*rand(1, N); % b is "tol" away from a
a_bin = floor(a/tol);
b_bin = floor(b/tol);
result = ismember(b_bin, a_bin) | ...
ismember(b_bin, a_bin-1) | ...
ismember(b_bin, a_bin+1);
find(result==0) % should be empty matrix.
The idea is to discretize the a and b variables to bins of size tol. Then, you ask whether b is found in the same bin as any element from a, or in the bin to the left of it, or in the bin to the right of it.
Advantages: I believe ismember is clever inside, first sorting the elements of a and then performing sublinear (log(N)) search per element b. This is unlike approaches which explicitly construct differences of each element in b with elements from a, meaning the complexity is linear in the number of elements in a.
Comparison: for N=100000 this runs 0.04s on my machine, compared to 20s using linear search (timed using Alan's nice and concise tf = arrayfun(#(bi) any(abs(a - bi) < tol), b); solution).
Disadvantages: this leads to that the actual tolerance is anything between tol and 1.5*tol. Depends on your task whether you can live with that (if the only concern is floating point comparison, you can).
Note: whether this is a viable approach depends on the ranges of a and b, and value of tol. If a and b can be very big and tol is very small, the a_bin and b_bin will not be able to resolve individual bins (then you would have to work with integral types, again checking carefully that their ranges suffice). The solution with loops is a safer one, but if you really need speed, you can invest into optimizing the presented idea. Another option, of course, would be to write a mex extension.
It sounds like what you are trying to do is have an ismember function for use on real valued data.
That is, check for each value B(i) in your vector B whether B(i) is within the tolerance threshold T of at least one value in your vector A
This works out something like the following:
tf = false(1, length(b)); %//the result vector, true if that element of b is in a
t = 0.01; %// the tolerance threshold
for i = 1:length(b)
%// is the absolute difference between the
%//element of a and b less that the threshold?
matches = abs(a - b(i)) < t;
%// if b(i) matches any of the elements of a
tf(i) = any(matches);
end
Or, in short:
t = 0.01;
tf = arrayfun(#(bi) any(abs(a - bi) < t), b);
Regarding avoiding the for loop: while this might benefit from vectorization, you may also want to consider looking at parallelisation if your data is that huge. In that case having a for loop as in my first example can be handy since you can easily do a basic version of parallel processing by changing the for to parfor.
Here is a fully vectorized solution. Note that I would actually recommend the solution given by #Alan, as mine is not likely to work for big datasets.
[X Y]=meshgrid(A,B)
M=abs(X-Y)<tolerance
Now the logical index of elements in a that are within the tolerance can be obtained with any(M) and the index for B is found by any(M,2)
bsxfun to the rescue
>> M = abs( bsxfun(#minus, A, B' ) ); %//' difference
>> M < tolerance
Another way to do what you want is with a logical expression.
Since A and B are vectors of different sizes you can't simply subtract and look for values that are smaller than the tolerance, but you can do the following:
Lmat = sparse((abs(repmat(A,[numel(B) 1])-repmat(B',[1 numel(A)])))<tolerance);
and you will get a sparse logical matrix with as many ones in it as equal elements (within tolerance). You could then count how many of those elements you have by writing:
Nequal = sum(sum(Lmat));
You could also get the indexes of the corresponding elements by writing:
[r,c] = find(Lmat);
then the following code will be true (for all j in numel(r)):
B(r(j))==A(c(j))
Finally, you should note that this way you get multiple counts in case there are duplicate entries in A or in B. It may be advisable to use the unique function first. For example:
A_new = unique(A);
I have 12 sets of vectors (about 10-20 vectors each) and i want to pick one vector of each set so that a function f that takes the sum of these vectors as argument is maximized. In addition i have constraints for some components of that sum.
Example:
a_1 = [3 2 0 5], a_2 = [3 0 0 2], a_3 = [6 0 1 1], ... , a_20 = [2 12 4 3]
b_1 = [4 0 4 -2], b_2 = [0 0 1 0], b_3 = [2 0 0 4], ... , b_16 = [0 9 2 3]
...
l_1 = [4 0 2 0], l_2 = [0 1 -2 0], l_3 = [4 4 0 1], ... , l_19 = [3 0 9 0]
s = [s_1 s_2 s_3 s_4] = a_x + b_y + ... + l_z
Constraints:
s_1 > 40
s_2 < 100
s_4 > -20
Target: Chose x, y, ... , z to maximize f(s):
f(s) -> max
Where f is a nonlinear function that takes the vector s and returns a scalar.
Bruteforcing takes too long because there are about 5.9 trillion combinations, and since i need the maximum (or even better the top 10 combinations) i can not use any of the greedy algorithms that came to my mind.
The vectors are quite sparse, about 70-90% are zeros. If that is helping somehow ...?
The Matlab Optimization toolbox didnt help either since it doesnt much support for discrete optimization.
Basically this is a lock-picking problem, where the lock's pins have 20 distinct positions, and there are 12 pins. Also:
some of the pin's positions will be blocked, depending on the positions of all the other pins.
Depending on the specifics of the lock, there may be multiple keys that fit
...interesting!
Based on Rasman's approach and Phpdna's comment, and the assumption that you are using int8 as data type, under the given constraints there are
>> d = double(intmax('int8'));
>> (d-40) * (d+100) * (d+20) * 2*d
ans =
737388162
possible vectors s (give or take a few, haven't thought about +1's etc.). ~740 million evaluations of your relatively simple f(s) shouldn't take more than 2 seconds, and having found all s that maximize f(s), you are left with the problem of finding linear combinations in your vector set that add up to one of those solutions s.
Of course, this finding of combinations is no easy feat, and the whole method breaks down anyway if you are dealing with
int16: ans = 2.311325368800510e+018
int32: ans = 4.253529737045237e+037
int64: ans = 1.447401115466452e+076
So, I'll discuss a more direct and more general approach here.
Since we're talking integers and a fairly large search space, I'd suggest using a branch-and-bound algorithm. But unlike the bintprog algorithm, you'd have to use different branching strategies, and of course, these should be based on a non-linear objective function.
Unfortunately, there is nothing like this in the optimization toolbox (or the File Exchange as far as I could find). fmincon is a no-go, since it uses gradient and Hessian information (which will usually be all-zero for integers), and fminsearch is a no-go, since you'll need a really good initial estimate, and the rate of convergence is (roughly) O(N), meaning, for this 20-dimensional problem you'll have to wait quite long before convergence, without the guarantee of having found the global solution.
An interval method could be a possibility, however, I personally have very little experience with this. There is no native interval-related stuff in MATLAB or any of its toolboxes, but there's the freely available INTLAB.
So, if you're not feeling like implementing your own non-linear binary integer programming algorithm, or are not in the mood for an adventure with INTLAB, there's really only one thing left: heuristic methods. In this link there is a similar situation, with an outline of the solution: use the genetic algorithm (ga) from the Global Optimization toolbox.
I would implement the problem roughly like so:
function [sol, fval, exitflag] = bintprog_nonlinear()
%// insert your data here
%// Any sparsity you may have here will only make this more
%// *memory* efficient, not *computationally*
data = [...
... %// this will be an array with size 4-by-20-by-12
... %// (or some permutation of that you find more intuitive)
];
%// offsets into the 3D array to facilitate indexing a bit
offsets = bsxfun(#plus, ...
repmat(1:size(data,1), size(data,3),1), ...
(0:size(data,3)-1)' * size(data,1)*size(data,2)); %//'
%// your objective function
function val = obj(X)
%// limit "X" to integers in [1 20]
X = min(max(round(X),1),size(data,3));
%// "X" will be a collection of 12 integers between 0 and 20, which are
%// indices into the data matrix
%// form "s" from "X"
s = sum(bsxfun(#plus, offsets, X*size(data,1) - size(data,1)));
%// XxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxX
%// Compute the NEGATIVE VALUE of your function here
%// XxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxXxX
end
%// your "non-linear" constraint function
function [C, Ceq] = nonlcon(X)
%// limit "X" to integers in [1 20]
X = min(max(round(X),1),size(data,3));
%// form "s" from "X"
s = sum(bsxfun(#plus, offsets, X(:)*size(data,1) - size(data,1)));
%// we have no equality constraints
Ceq = [];
%// Compute inequality constraints
%// NOTE: solver is trying to solve C <= 0, so:
C = [...
40 - s(1)
s(2) - 100
-20 - s(4)
];
end
%// useful GA options
options = gaoptimset(...
'UseParallel', 'always'...
...
);
%// The rest really depends on the specifics of the problem.
%// Useful to look at will be at least 'TolCon', 'Vectorized', and of course,
%// 'PopulationType', 'Generations', etc.
%// THE OPTIMZIATION
[sol, fval, exitflag] = ga(...
#obj, size(data,3), ... %// objective function, taking a vector of 20 values
[],[], [],[], ... %// no linear (in)equality constraints
1,size(data,2), ... %// lower and upper limits
#nonlcon, options); %// your "nonlinear" constraints
end
Note that even though your constraints are essentially linear, the way by which you must compute the value for your s necessitates the use of a custom constraint function (nonlcon).
Especially note that this is currently (probably) a sub-optimal way to use ga -- I don't know the specifics of your objective function, so a lot more may be possible. For instance, I currently use a simple round() to convert the input X to integers, but using 'PopulationType', 'custom' (with a custom 'CreationFcn', 'MutationFcn' etc.) might produce better results. Also, 'Vectorized' will likely speed things up a lot, but I don't know whether your function is easily vectorized.
And yes, I use nested functions (I just love those things!); it prevents these huge, usually identical lists of input arguments if you use sub-functions or stand-alone functions, and they can really be a performance boost because there is little copying of data. But, I realize that their scoping rules make them somewhat akin to goto constructs, and so they are -ahum- "not everyone's cup of tea"...you might want to convert them to sub-functions to prevent long and useless discussions with your co-workers :)
Anyway, this should be a good place to start. Let me know if this is useful at all.
Unless you define some intelligence on how the vector sets are organized, there will be no intelligent way of solving your problem other then pure brute force.
Say you find s s.t. f(s) is max given constraints of s, you still need to figure out how to build s with twelve 4-element vectors (an overdetermined system if there ever was one), where each vector has 20 possible values. Sparsity may help, although I'm not sure how it is possible to have a vector with four elements be 70-90% zero, and sparsity would only be useful if there was some yet to be described methodology in how the vector are organized
So I'm not saying you can't solve the problem, I'm saying you need to rethink how the problem is set-up.
I know, this answer is reaching you really late.
Unfortunately, the problem, as is, show not many patterns to be exploited, besides of brute force -Branch&Bound, Master& Slave, etc.- Trying a Master Slave approach -i.e. solving first the function continuous nonlinear problem as master, and solving the discrete selection as slave could help, but with as many combinations, and without any more information over the vectors, there is not too much space for work.
But based on the given continuous almost everywhere functions, based on combinations of sums and multiplication operators and their inverses, the sparsity is a clear point to be exploited here. If 70-90% of vectors are zero, almost a good part of the solution space will be close to zero, or close to infinite. Hence a 80-20 pseudo solution would discard easily the 'zero' combinations, and use only the 'infinite' ones.
This way, the brute-force could be guided.
I'm having trouble creating a random vector V in Matlab subject to the following set of constraints: (given parameters N,D, L, and theta)
The vector V must be N units long
The elements must have an average of theta
No 2 successive elements may differ by more than +/-10
D == sum(L*cosd(V-theta))
I'm having the most problems with the last one. Any ideas?
Edit
Solutions in other languages or equation form are equally acceptable. Matlab is just a convenient prototyping tool for me, but the final algorithm will be in java.
Edit
From the comments and initial answers I want to add some clarifications and initial thoughts.
I am not seeking a 'truly random' solution from any standard distribution. I want a pseudo randomly generated sequence of values that satisfy the constraints given a parameter set.
The system I'm trying to approximate is a chain of N links of link length L where the end of the chain is D away from the other end in the direction of theta.
My initial insight here is that theta can be removed from consideration until the end, since (2) in essence adds theta to every element of a 0 mean vector V (shifting the mean to theta) and (4) simply removes that mean again. So, if you can find a solution for theta=0, the problem is solved for all theta.
As requested, here is a reasonable range of parameters (not hard constraints, but typical values):
5<N<200
3<D<150
L==1
0 < theta < 360
I would start by creating a "valid" vector. That should be possible - say calculate it for every entry to have the same value.
Once you got that vector I would apply some transformations to "shuffle" it. "Rejection sampling" is the keyword - if the shuffle would violate one of your rules you just don't do it.
As transformations I come up with:
switch two entries
modify the value of one entry and modify a second one to keep the 4th condition (Theoretically you could just shuffle two till the condition is fulfilled - but the chance that happens is quite low)
But maybe you can find some more.
Do this reasonable often and you get a "valid" random vector. Theoretically you should be able to get all valid vectors - practically you could try to construct several "start" vectors so it won't take that long.
Here's a way of doing it. It is clear that not all combinations of theta, N, L and D are valid. It is also clear that you're trying to simulate random objects that are quite complex. You will probably have a hard time showing anything useful with respect to these vectors.
The series you're trying to simulate seems similar to the Wiener process. So I started with that, you can start with anything that is random yet reasonable. I then use that as a starting point for an optimization that tries to satisfy 2,3 and 4. The closer your initial value to a valid vector (satisfying all your conditions) the better the convergence.
function series = generate_series(D, L, N,theta)
s(1) = theta;
for i=2:N,
s(i) = s(i-1) + randn(1,1);
end
f = #(x)objective(x,D,L,N,theta)
q = optimset('Display','iter','TolFun',1e-10,'MaxFunEvals',Inf,'MaxIter',Inf)
[sf,val] = fminunc(f,s,q);
val
series = sf;
function value= objective(s,D,L,N,theta)
a = abs(mean(s)-theta);
b = abs(D-sum(L*cos(s-theta)));
c = 0;
for i=2:N,
u =abs(s(i)-s(i-1)) ;
if u>10,
c = c + u;
end
end
value = a^2 + b^2+ c^2;
It seems like you're trying to simulate something very complex/strange (a path of a given curvature?), see questions by other commenters. Still you will have to use your domain knowledge to connect D and L with a reasonable mu and sigma for the Wiener to act as initialization.
So based on your new requirements, it seems like what you're actually looking for is an ordered list of random angles, with a maximum change in angle of 10 degrees (which I first convert to radians), such that the distance and direction from start to end and link length and number of links are specified?
Simulate an initial guess. It will not hold with the D and theta constraints (i.e. specified D and specified theta)
angles = zeros(N, 1)
for link = 2:N
angles (link) = theta(link - 1) + (rand() - 0.5)*(10*pi/180)
end
Use genetic algorithm (or another optimization) to adjust the angles based on the following cost function:
dx = sum(L*cos(angle));
dy = sum(L*sin(angle));
D = sqrt(dx^2 + dy^2);
theta = atan2(dy/dx);
the cost is now just the difference between the vector given by my D and theta above and the vector given by the specified D and theta (i.e. the inputs).
You will still have to enforce the max change of 10 degrees rule, perhaps that should just make the cost function enormous if it is violated? Perhaps there is a cleaner way to specify sequence constraints in optimization algorithms (I don't know how).
I feel like if you can find the right optimization with the right parameters this should be able to simulate your problem.
You don't give us a lot of detail to work with, so I'll assume the following:
random numbers are to be drawn from [-127+theta +127-theta]
all random numbers will be drawn from a uniform distribution
all random numbers will be of type int8
Then, for the first 3 requirements, you can use this:
N = 1e4;
theta = 40;
diffVal = 10;
g = #() randi([intmin('int8')+theta intmax('int8')-theta], 'int8') + theta;
V = [g(); zeros(N-1,1, 'int8')];
for ii = 2:N
V(ii) = g();
while abs(V(ii)-V(ii-1)) >= diffVal
V(ii) = g();
end
end
inline the anonymous function for more speed.
Now, the last requirement,
D == sum(L*cos(V-theta))
is a bit of a strange one...cos(V-theta) is a specific way to re-scale the data to the [-1 +1] interval, which the multiplication with L will then scale to [-L +L]. On first sight, you'd expect the sum to average out to 0.
However, the expected value of cos(x) when x is a random variable from a uniform distribution in [0 2*pi] is 2/pi (see here for example). Ignoring for the moment the fact that our limits are different from [0 2*pi], the expected value of sum(L*cos(V-theta)) would simply reduce to the constant value of 2*N*L/pi.
How you can force this to equal some other constant D is beyond me...can you perhaps elaborate on that a bit more?
I have a 101x82 size matrix called A. Using this variable matrix, I compute two other variables called:
1) B, a 1x1 scalar, and
2) C, a 50x6 matrix.
I compare 1) and 2) with their analogues variables 3) and 4), whose values are fixed:
3) D, a 1x1 scalar, and
4) E, a 50x6 matrix.
Now, I want to perturb/change the values of A matrix, such that:
1) ~ 3), i.e. B is nearly equal to D , and
2) ~ 4), i.e. C is nearly equal to E
Note that on perturbing A, B and C will change, but not D and E.
Any ideas how to do this? Thanks!
I can't run your code as it's demanding to load data (which I don't have) and it's not immediatly obvious how to calculate B or C.
Fortunately I may be able to answer your problem. You're describing an optimization problem, and the solution would be to use fminsearch (or something of that variety).
What you do is define a function that returns a vector with two elements:
y1 = (B - D)^weight1;
y2 = norm((C - E), weight2);
with weight being how strong you allow for variability (weight = 2 is usually sufficient).
Your function variable would be A.
From my understanding you have a few functions.
fb(A) = B
fc(A) = C
Do you know the functions listed above, that is do you know the mappings from A to each of these?
If you want to try to optimize, so that B is close to D, you need to pick:
What close means. You can look at some vector norm for the B and D case, like minimizing ||B-D||^2. The standard sum of the squares of the elements of this different will probably do the trick and is computationally nice.
How to optimize. This depends a lot on your functions, whether you want local or global mimina, etc.
So basically, now we've boiled the problem down to minimizing:
Cost = ||fb(A) - fd(A)||^2
One thing you can certainly do is to compute the gradient of this cost function with respect to the individual elements of A, and then perform minimization steps with forward Euler method with a suitable "time step". This might not be fast, but with small enough time step and well-behaved enough functions it will at least get you to a local minima.
Computing the gradient of this
grad_A(cost) = 2*||fb(A)-fd(A)||*(grad_A(fb)(A)-grad_A(fd)(A))
Where grad_A means gradient with respect to A, and grad_A(fb)(A) means gradient with respect to A of the function fb evaluated at A, etc.
Computing the grad_A(fb)(A) depends on the form of fb, but here are some pages have "Matrix calculus" identities and explanations.
Matrix calculus identities
Matrix calculus explanation
Then you simply perform gradient descent on A by doing forward Euler updates:
A_next = A_prev - timestep * grad_A(cost)