I am attempting a maximisation problem subject to various constraints.
i.e. max y = x1 + x2 + x3 + .... + xn
where each xi is a vector of values over time: x1 = (x11, x12, x13,...)
Some of the constraints state that specific values of xit cannot be positive in the same time period.
i.e. if(x1t > 0), x2t = 0; if(x2t > 0), x1t = 0
For context, the constraint is equivalent to "maximise the revenue of a shop, but you cant sell product A and B on the same day"
How do I go about formulating an LP model in Excel (using solver) to solve this.
This is called a complementarity constraint. One way of modeling this is:
x(1,t) * x(2,t) = 0
x(i,t) ≥ 0
However, this is nonlinear (and in a somewhat nasty way). A linear approach, using an extra binary variable δ can look like:
x(1,t) ≤ UP(1,t) * δ(t)
x(2,t) ≤ UP(2,t) * (1-δ(t))
x(i,t) ∈ [0,UP(i,t)] 'UP is an upper bound on x'
δ(t) ∈ {0,1} 'δ is a binary variable'
Related
How can we calculate the correlation and covariance between two variables without using cov and corr in Python3?
At the end, I want to write a function that returns three values:
a boolean that is true if two variables are independent
covariance of two variables
correlation of two variables.
You can find the definition of correlation and covariance here:
https://medium.com/analytics-vidhya/covariance-and-correlation-math-and-python-code-7cbef556baed
I wrote this part for covariance:
'''
ans=[]
mean_x , mean_y = x.mean() , y.mean()
n = len(x)
Cov = sum((x - mean_x) * (y - mean_y)) / n
sum_x = float(sum(x))
sum_y = float(sum(y))
sum_x_sq = sum(xi*xi for xi in x)
sum_y_sq = sum(yi*yi for yi in y)
psum = sum(xi*yi for xi, yi in zip(x, y))
num = psum - (sum_x * sum_y/n)
den = pow((sum_x_sq - pow(sum_x, 2) / n) * (sum_y_sq - pow(sum_y, 2) / n), 0.5)
if den == 0: return 0
return num / den
'''
For the covariance, just subtract the respective means and multiply the vectors together (using the dot product). (Of course, make sure whether you're using the sample covariance or population covariance estimate -- if you have "enough" data the difference will be tiny, but you should still account for it if necessary.)
For the correlation, divide the covariance by the standard deviations of both.
As for whether or not two columns are independent, that's not quite as easy. For two random variables, we just have that $\mathbb{E}\left[(X - \mu_X)(Y - \mu_Y)\right] = 0$, where $\mu_X, \mu_Y$ are the means of the two variables. But, when you have a data set, you are not dealing with the actual probability distributions; you are dealing with a sample. That means that the correlation will very likely not be exactly $0$, but rather a value close to $0$. Whether or not this is "close enough" will depend on your sample size and what other assumptions you're willing to make.
I have a custom (discrete) probability distribution defined somewhat in the form: f(x)/(sum(f(x')) for x' in a given discrete set X). Also, 0<=x<=1.
So I have been trying to implement it in python 3.8.2, and the problem is that the numerator and denominator both come out to be really small and python's floating point representation just takes them as 0.0.
After calculating these probabilities, I need to sample a random element from an array, whose each index may be selected with the corresponding probability in the distribution. So if my distribution is [p1,p2,p3,p4], and my array is [a1,a2,a3,a4], then probability of selecting a2 is p2 and so on.
So how can I implement this in an elegant and efficient way?
Is there any way I could use the np.random.beta() in this case? Since the difference between the beta distribution and my actual distribution is only that the normalization constant differs and the domain is restricted to a few points.
Note: The Probability Mass function defined above is actually in the form given by the Bayes theorem and f(x)=x^s*(1-x)^f, where s and f are fixed numbers for a given iteration. So the exact problem is that, when s or f become really large, this thing goes to 0.
You could well compute things by working with logs. The point is that while both the numerator and denominator might underflow to 0, their logs won't unless your numbers are really astonishingly small.
You say
f(x) = x^s*(1-x)^t
so
logf (x) = s*log(x) + t*log(1-x)
and you want to compute, say
p = f(x) / Sum{ y in X | f(y)}
so
p = exp( logf(x) - log sum { y in X | f(y)}
= exp( logf(x) - log sum { y in X | exp( logf( y))}
The only difficulty is in computing the second term, but this is a common problem, for example here
On the other hand computing logsumexp is easy enough to to by hand.
We want
S = log( sum{ i | exp(l[i])})
if L is the maximum of the l[i] then
S = log( exp(L)*sum{ i | exp(l[i]-L)})
= L + log( sum{ i | exp( l[i]-L)})
The last sum can be computed as written, because each term is now between 0 and 1 so there is no danger of overflow, and one of the terms (the one for which l[i]==L) is 1, and so if other terms underflow, that is harmless.
This may however lose a little accuracy. A refinement would be to recognize the set A of indices where
l[i]>=L-eps (eps a user set parameter, eg 1)
And then compute
N = Sum{ i in A | exp(l[i]-L)}
B = log1p( Sum{ i not in A | exp(l[i]-L)}/N)
S = L + log( N) + B
I constructed a finite automata for the language L of all strings made of the symbols 0, 1 and 2 (Σ = {0, 1, 2}) where the last symbol is not smaller than the first symbol. E.g., the strings 0, 2012, 01231 and 102 are in the language, but 10, 2021 and 201 are not in the language.
Then from that an GNFA so I can convert to RE.
My RE looks like this:
(0(0+1+2)* )(1(0(1+2)+1+2)* )(2((0+1)2+2))*)
I have no idea if this is correct, as I think I understand RE but not entirely sure.
Could someone please tell me if it’s correct and if not why?
There is a general method to convert any DFA into a regular expression, and is probably what you should be using to solve this homework problem.
For your attempt specifically, you can tell whether an RE is incorrect by finding a word that should be in the language, but that your RE doesn't accept, or a word that shouldn't be in the language that the RE does accept. In this case, the string 1002 should be in the language, but the RE doesn't match it.
There are two primary reasons why this string isn't matched. The first is that there should be a union rather than a concatenation between the three major parts of the language (words starting with 0, 1 and 2, respectively:
(0(0+1+2)*) (1(0(1+2)+1+2)*) (2((0+1)2+2))*) // wrong
(0(0+1+2)*) + (1(0(1+2)+1+2)*) + (2((0+1)2+2))*) // better
The second problem is that in the 1 and 2 cases, the digits smaller than the starting digit need to be repeatable:
(1(0 (1+2)+1+2)*) // wrong
(1(0*(1+2)+1+2)*) // better
If you do both of those things, the RE will be correct. I'll leave it as an exercise for you to follow that step for the 2 case.
The next thing you can try is find a way to make the RE more compact:
(1(0*(1+2)+1+2)*) // verbose
(1(0*(1+2))*) // equivalent, but more compact
This last step is just a matter of preference. You don't need the trailing +1+2 because 0* can be of zero length, so 0*(1+2) covers the +1+2 case.
You can use an algorithm but this DFA might be easy enough to convert as a one-off.
First, note that if the first symbol seen in the initial state is 0, you transition to state A and remain there. A is accepting. This means any string beginning with 0 is accepted. Thus, our regular expression might as well have a term like 0(0+1+2)* in it.
Second, note that if the first symbol seen in the initial state is 1, you transition to state B and remain in states B and D from that point on. You only leave B if you see 0 and you stay out of B as long as you keep seeing 0. The only way to end on D is if the last symbol you saw was 0. Therefore, strings beginning with 1 are accepted if and only if the strings don't end in 0. We can have a term like 1(0+1+2)*(1+2) in our regular expression as well to cover these cases.
Third, note that if the first symbol seen in the initial state is 2, you transition to state C and remain in states C and E from that point on. You leave state C if you see anything but 2 and stay out of B until you see a 2 again. The only way to end up on C is if the last symbol you saw was 2. Therefore, strings beginning with 2 are accepted if and only if the strings end in 2. We can have a term like 2(0+1+2)*(2) in our regular expression as well to cover these cases.
Finally, we see that there are no other cases to consider; our three terms cover all cases and the union of them fully describes our language:
0(0+1+2)* + 1(0+1+2)*(1+2) + 2(0+1+2)*2
It was easy to just write out the answer here because this DFA is sort of like three simple DFAs put together with a start state. More complicated DFAs might be easier to convert to REs using algorithms that don't require you understand or follow what the DFA is doing.
Note that if the start state is accepting (mentioned in a comment on another answer) the RE changes as follows:
e + 0(0+1+2)* + 1(0+1+2)*(1+2) + 2(0+1+2)*2
Basically, we just tack the empty string onto it since it is not already generated by any of the other parts of the aggregate expression.
You have the equivalent of what is known as a right-linear system. It's right-linear because the variables occur on the right hand sides only to the first degree and only on the right-hand sides of each term. The system that you have may be written - with a change in labels from 0,1,2 to u,v,w - as
S ≥ u A + v B + w C
A ≥ 1 + (u + v + w) A
B ≥ 1 + u D + (v + w) B
C ≥ 1 + (u + v) E + w C
D ≥ u D + (v + w) B
E ≥ (u + v) E + w C
The underlying algebra is known as a Kleene algebra. It is defined by the following identities that serve as its fundamental properties
(xy)z = x(yz), x1 = x = 1x,
(x + y) + z = x + (y + z), x + 0 = x = 0 + x,
y0z = 0, w(x + y)z = wxz + wyz,
x + y = y + x, x + x = x,
with a partial ordering relation defined by
x ≤ y ⇔ y ≥ x ⇔ ∃z(x + z = y) ⇔ x + y = y
With respect to this ordering relation, all finite subsets have least upper bounds, including the following
0 = ⋁ ∅, x + y = ⋁ {x, y}
The sum operator "+" is the least upper bound operator.
The system you have is a right-linear fixed point system, since it expresses the variables on the left as a (right-linear) function, as given on the right, of the variables. The object being specified by the system is the least solution with respect to the ordering; i.e. the least fixed point solution; and the regular expression sought out is the value that the main variable has in the least fixed point solution.
The last axiom(s) for Kleene algebras can be stated in any of a number of equivalent ways, including the following:
0* = 1
the least fixed point solution to x ≥ a + bx + xc is x = b* a c*.
There are other ways to express it. A consequence is that one has identities such as the following:
1 + a a* = a* = 1 + a* a
(a + b)* = a* (b a*)*
(a b)* a = a (b a)*
In general, right linear systems, such as the one corresponding to your problem may be written in vector-matrix form as 𝐪 ≥ 𝐚 + A 𝐪, with the least fixed point solution given in matrix form as 𝐪 = A* 𝐚. The central theorem of Kleene algebras is that all finite right-linear systems have least fixed point solutions; so that one can actually define matrix algebras over Kleene algebras with product and sum given respectively as matrix product and matrix sum, and that this algebra can be made into a Kleene algebra with a suitably-defined matrix star operation through which the least fixed point solution is expressed. If the matrix A decomposes into block form as
B C
D E
then the star A* of the matrix has the block form
(B + C E* D)* (B + C E* D)* C E*
(E + D B* C)* D B* (E + D B* C)*
So, what this is actually saying is that for a vector-matrix system of the form
x ≥ a + B x + C y
y ≥ b + D x + E y
the least fixed point solution is given by
x = (B + C E* D)* (a + C E* b)
y = (E + D B* C)* (D B* a + b)
The star of a matrix, if expressed directly in terms of its components, will generally be huge and highly redundant. For an n×n matrix, it has size O(n³) - cubic in n - if you allow for redundant sub-expressions to be defined by macros. Otherwise, if you in-line insert all the redundancy then I think it blows up to a highly-redundant mess that is exponential in n in size.
So, there's intelligence required and involved (literally meaning: AI) in finding or pruning optimal forms that avoid the blow-up as much as possible. That's a non-trivial job for any purported matrix solver and regular expression synthesis compiler.
An heuristic, for your system, is to solve for the variables that don't have a "1" on the right-hand side and in-line substitute the solutions - and to work from bottom-up in terms of the dependency chain of the variables. That would mean starting with D and E first
D ≥ u* (v + w) B
E ≥ (u + v)* w C
In-line substitute into the other inequations
S ≥ u A + v B + w C
A ≥ 1 + (u + v + w) A
B ≥ 1 + u u* (v + w) B + (v + w) B
C ≥ 1 + (u + v) (u + v)* w C + w C
Apply Kleene algebra identities (e.g. x x* y + y = x* y)
S ≥ u A + v B + w C
A ≥ 1 + (u + v + w) A
B ≥ 1 + u* (v + w) B
C ≥ 1 + (u + v)* w C
Solve for the next layer of dependency up: A, B and C:
A ≥ (u + v + w)*
B ≥ (u* (v + w))*
C ≥ ((u + v)* w)*
Apply some more Kleene algebra (e.g. (x* y)* = 1 + (x + y)* y) to get
B ≥ 1 + N (v + w)
C ≥ 1 + N w
where, for convenience we set N = (u + v + w)*. In-line substitute at the top-level:
S ≥ u N + v (1 + N (v + w)) + w (1 + N w).
The least fixed point solution, in the main variable S, is thus:
S = u N + v + v N (v + w) + w + w N w.
where
N = (u + v + w)*.
As you can already see, even with this simple example, there's a lot of chess-playing to navigate through the system to find an optimally-pruned solution. So, it's certainly not a trivial problem. What you're essentially doing is synthesizing a control-flow structure for a program in a structured programming language from a set of goto's ... essentially the core process of reverse-compiling from assembly language to a high level language.
One measure of optimization is that of minimizing the loop-depth - which here means minimizing the depth of the stars or the star height. For example, the expression x* (y x*)* has star-height 2 but reduces to (x + y)*, which has star height 1. Methods for reducing star-height come out of the research by Hashiguchi and his resolution of the minimal star-height problem. His proof and solution (dating, I believe, from the 1980's or 1990's) is complex and to this day the process still goes on of making something more practical of it and rendering it in more accessible form.
Hashiguchi's formulation was cast in the older 1950's and 1960's formulation, predating the axiomatization of Kleene algebras (which was in the 1990's), so to date, nobody has rewritten his solution in entirely algebraic form within the framework of Kleene algebras anywhere in the literature ... as far as I'm aware. Whoever accomplishes this will have, as a result, a core element of an intelligent regular expression synthesis compiler, but also of a reverse-compiler and programming language synthesis de-compiler. Essentially, with something like that on hand, you'd be able to read code straight from binary and the lid will be blown off the world of proprietary systems. [Bite tongue, bite tongue, mustn't reveal secret yet, must keep the ring hidden.]
I'm attempting to solve the differential equation:
m(t) = M(x)x'' + C(x, x') + B x'
where x and x' are vectors with 2 entries representing the angles and angular velocity in a dynamical system. M(x) is a 2x2 matrix that is a function of the components of theta, C is a 2x1 vector that is a function of theta and theta' and B is a 2x2 matrix of constants. m(t) is a 2*1001 array containing the torques applied to each of the two joints at the 1001 time steps and I would like to calculate the evolution of the angles as a function of those 1001 time steps.
I've transformed it to standard form such that :
x'' = M(x)^-1 (m(t) - C(x, x') - B x')
Then substituting y_1 = x and y_2 = x' gives the first order linear system of equations:
y_2 = y_1'
y_2' = M(y_1)^-1 (m(t) - C(y_1, y_2) - B y_2)
(I've used theta and phi in my code for x and y)
def joint_angles(theta_array, t, torques, B):
phi_1 = np.array([theta_array[0], theta_array[1]])
phi_2 = np.array([theta_array[2], theta_array[3]])
def M_func(phi):
M = np.array([[a_1+2.*a_2*np.cos(phi[1]), a_3+a_2*np.cos(phi[1])],[a_3+a_2*np.cos(phi[1]), a_3]])
return np.linalg.inv(M)
def C_func(phi, phi_dot):
return a_2 * np.sin(phi[1]) * np.array([-phi_dot[1] * (2. * phi_dot[0] + phi_dot[1]), phi_dot[0]**2])
dphi_2dt = M_func(phi_1) # (torques[:, t] - C_func(phi_1, phi_2) - B # phi_2)
return dphi_2dt, phi_2
t = np.linspace(0,1,1001)
initial = theta_init[0], theta_init[1], dtheta_init[0], dtheta_init[1]
x = odeint(joint_angles, initial, t, args = (torque_array, B))
I get the error that I cannot index into torques using the t array, which makes perfect sense, however I am not sure how to have it use the current value of the torques at each time step.
I also tried putting odeint command in a for loop and only evaluating it at one time step at a time, using the solution of the function as the initial conditions for the next loop, however the function simply returned the initial conditions, meaning every loop was identical. This leads me to suspect I've made a mistake in my implementation of the standard form but I can't work out what it is. It would be preferable however to not have to call the odeint solver in a for loop every time, and rather do it all as one.
If helpful, my initial conditions and constant values are:
theta_init = np.array([10*np.pi/180, 143.54*np.pi/180])
dtheta_init = np.array([0, 0])
L_1 = 0.3
L_2 = 0.33
I_1 = 0.025
I_2 = 0.045
M_1 = 1.4
M_2 = 1.0
D_2 = 0.16
a_1 = I_1+I_2+M_2*(L_1**2)
a_2 = M_2*L_1*D_2
a_3 = I_2
Thanks for helping!
The solver uses an internal stepping that is problem adapted. The given time list is a list of points where the internal solution gets interpolated for output samples. The internal and external time lists are in no way related, the internal list only depends on the given tolerances.
There is no actual natural relation between array indices and sample times.
The translation of a given time into an index and construction of a sample value from the surrounding table entries is called interpolation (by a piecewise polynomial function).
Torque as a physical phenomenon is at least continuous, a piecewise linear interpolation is the easiest way to transform the given function value table into an actual continuous function. Of course one also needs the time array.
So use numpy.interp1d or the more advanced routines of scipy.interpolate to define the torque function that can be evaluated at arbitrary times as demanded by the solver and its integration method.
I found this code snippet on raywenderlich.com, however the link to the explanation wasn't valid anymore. I "translated" the answer into Swift, I hope you can understand, it's actually quite easy even without knowing the language. Could anyone explain what exactly is going on here? Thanks for any help.
class func linesCross(#line1: Line, line2: Line) -> Bool {
let denominator = (line1.end.y - line1.start.y) * (line2.end.x - line2.start.x) -
(line1.end.x - line1.start.x) * (line2.end.y - line2.start.y)
if denominator == 0 { return false } //lines are parallel
let ua = ((line1.end.x - line1.start.x) * (line2.start.y - line1.start.y) -
(line1.end.y - line1.start.y) * (line2.start.x - line1.start.x)) / denominator
let ub = ((line2.end.x - line2.start.x) * (line2.start.y - line1.start.y) -
(line2.end.y - line2.start.y) * (line2.start.x - line1.start.x)) / denominator
//lines may touch each other - no test for equality here
return ua > 0 && ua < 1 && ub > 0 && ub < 1
}
You can find a detailed segment-intersection algorithm
in the book Computational Geometry in C, Sec. 7.7.
The SegSegInt code described there is available here.
I recommend avoiding slope calculations.
There are several "degenerate" cases that require care: collinear segments
overlapping or not, one segment endpoint in the interior of the other segments,
etc. I wrote the code to return an indication of these special cases.
This is what the code is doing.
Every point P in the segment AB can be described as:
P = A + u(B - A)
for some constant 0 <= u <= 1. In fact, when u=0 you get P=A, and you getP=B when u=1. Intermediate values of u will give you intermediate values of P in the segment. For instance, when u = 0.5 you will get the point in the middle. In general, you can think of the parameter u as the ratio between the lengths of AP and AB.
Now, if you have another segment CD you can describe the points Q on it in the same way, but with a different u, which I will call v:
Q = C + v(D - C)
Again, keep in mind that Q lies between C and D if, and only if, 0 <= v <= 1 (same as above for P).
To find the intersection between the two segments you have to equate P=Q. In other words, you need to find u and v, both between 0 and 1 such that:
A + u(B - A) = C + v(D - C)
So, you have this equation and you have to see if it is solvable within the given constraints on u and v.
Given that A, B, C and D are points with two coordinates x,y each, you can open the equation above into two equations:
ax + u(bx - ax) = cx + v(dx - cx)
ay + u(by - ay) = cy + v(dy - cy)
where ax = A.x, ay = A.y, etc., are the coordinates of the points.
Now we are left with a 2x2 linear system. In matrix form:
|bx-ax cx-dx| |u| = |cx-ax|
|by-ay cy-dy| |v| |cy-ay|
The determinant of the matrix is
det = (bx-ax)(cy-dy) - (by-ay)(cx-dx)
This quantity corresponds to the denominator of the code snippet (please check).
Now, multiplying both sides by the cofactor matrix:
|cy-dy dx-cx|
|ay-by bx-ax|
we get
det*u = (cy-dy)(cx-ax) + (dx-cx)(cy-ay)
det*v = (ay-by)(cx-ax) + (bx-ax)(cy-ay)
which correspond to the variables ua and ub defined in the code (check this too!)
Finally, once you have u and v you can check whether they are both between 0 and 1 and in that case return that there is intersection. Otherwise, there isn't.
For a given line the slope is
m=(y_end-y_start)/(x_end-x_start)
if two slopes are equal, the lines are parallel
m1=m1
(y1_end-y_start)/(x1_end-x1_start)=(y2_end-y2_start)/(x2_end-x2_start)
And this is equivalent to checking that the denominator is not zero,
Regarding the rest of the code, find the explanation on wikipedia under "Given two points on each line"