Carl Love

Carl Love

27666 Reputation

25 Badges

12 years, 131 days
Himself
Wayland, Massachusetts, United States
My name was formerly Carl Devore.

MaplePrimes Activity


These are replies submitted by Carl Love

@Markiyan Hirnyk 

There are only two complete sentences on the web page in question, and I cut-and-paste them right here: "Maple T.A. is the only testing and assessment system designed for courses involving mathematics. In this video, you’ll see 10 reasons why Maple T.A. is the best choice for your math, engineering, and science courses."  I am not taking sides on this issue, as I know nothing about WebAssign, and very little about MapleTA. I'm just pointing out that they clearly say "only", as was claimed.

@Markiyan Hirnyk 

Ah, I see. Of course the solution can only be unique upto a common multiple of numerator and denominator.

@Markiyan Hirnyk 

Ah, I see. Of course the solution can only be unique upto a common multiple of numerator and denominator.

@Markiyan Hirnyk

I don't understand the point you are trying to make in "One more difficulty". Why did you make eq3 and eq4 the same?

Also, it would useful to sparsify and block-diagonalize the system of linear equations (especially since an exact  solution is required, rather than floating point) by choosing 0 as an evaluation point as much as possible.

@Markiyan Hirnyk

I don't understand the point you are trying to make in "One more difficulty". Why did you make eq3 and eq4 the same?

Also, it would useful to sparsify and block-diagonalize the system of linear equations (especially since an exact  solution is required, rather than floating point) by choosing 0 as an evaluation point as much as possible.

Something that is suspicious to me (and this may be the impetus which caused the OP to originally question the final result) is that the value 1307, which appears in the definition of B and as the upper limit of integration, is suspiciously close to, but not equal to, the value of A at which the abs would make a difference; which you computed to be 1309.46. Replacing both 1307s with this makes a significant difference (rel. err. = 1.6%) in the final result.

Something that is suspicious to me (and this may be the impetus which caused the OP to originally question the final result) is that the value 1307, which appears in the definition of B and as the upper limit of integration, is suspiciously close to, but not equal to, the value of A at which the abs would make a difference; which you computed to be 1309.46. Replacing both 1307s with this makes a significant difference (rel. err. = 1.6%) in the final result.

When you say "their power is limited ... [to], for instance, no more than 3", do you mean that the total degree of the polynomials is at most 3? Or do you mean that the degree, or power, of each individual variable is at most 3? 

What do you mean by "solve"? Do you mean "approximate"? What is the "it" about whose truth you ask?

You've given some Maple commands, which work correctly. But what's the point of only using three terms of your series? Why aren't you satisfied with the output?

@amrramadaneg Well, if "the original problem is still found," how about answering the other questions so that we can help you?

To address directly the OP's issues:

mike_a wrote:

> I noticed some strange things with the list/table carrying my complex points. If I ask for individual values, like T[20], it won't give the correct value, but if I ask for a range, like T[1..20], it will display the correct values.

You'll have to post some code so we can figure out what's going on. But I'll tell you right off that your problem has nothing to do with the size of your data set.

> I get an error when trying to use complexplot saying either invalid range, if I don't have a range for the 2nd and 3rd argument, or else Error, (in plot) procedure expected, as range contains no plotting variable. I assume that there's some problem with the table being too large and not a problem with the complexplot command

No, it has nothing to do with the table being too large. But you'll need to post some code so that we can figure out the actual problem.

First work out these indexing problems on a much smaller version of your plot before moving on to the million-point plot.

To summarize what you said, for those who may have trouble reading it: You are trying to solve, using a least-squares method, a linear system A.X = B, where A has dimensions 1201 x 800 and B is a column vector. You have a tried an ad hoc method that involves explicitly inverting an 800 x 800 matrix, but ran out of memory.

You wrote:

> A^T*A*X=A^T*X.  let M=A^T*A.  M*X=A^T*X  finally  X=M^-1 * A^T*X

That's supposed to be B, not X, on the right sides of those equations! I hope that was just a typo in your post and not in your Maple code!

Several questions before we go any further:

  1. Is A a hardware floating point matrix?
  2. What percentage of the elements of A are non-zero? A rough estimate is fine.
  3. Have you tried LinearAlgebra:-LeastSquares?
  4. Can you make a smaller version of this problem that we can practice with before doing the full 1201 x 800?

A plot shows that Preben's one-term approximation is excellent even for x close to -2, and the original is visually indistinguishable from the approximation for x < -3.2. The plot below uses A and res as defined in Preben's code. If Digits, numpoints, and adaptive are not reduced from their default values (as I have done below), Maple will take a very long time to do the numerical integrations.

Digits:= 6:
plot([op(1,res), A], x= -3.3..-2.1, numpoints= 50, adaptive= 4);

I had originally plotted down to x = -10, but then it difficult to even see that there are two curves because they are so close.

A plot shows that Preben's one-term approximation is excellent even for x close to -2, and the original is visually indistinguishable from the approximation for x < -3.2. The plot below uses A and res as defined in Preben's code. If Digits, numpoints, and adaptive are not reduced from their default values (as I have done below), Maple will take a very long time to do the numerical integrations.

Digits:= 6:
plot([op(1,res), A], x= -3.3..-2.1, numpoints= 50, adaptive= 4);

I had originally plotted down to x = -10, but then it difficult to even see that there are two curves because they are so close.

Are you sure that you didn't switch n and m in your big-O formula? Also, do you mean to compute separately the GCDs of m pairs of numbers? or to compute the overall GCD of a set of m numbers?

First 691 692 693 694 695 696 697 Last Page 693 of 704