InverseSolve

InverseSolve - InverseSolve - 消息

#1 已发布: 2016/10/18 19:16:10
Jean Giraud

Jean Giraud

983个赞,共6,866个帖子。

组: User

Radovan:

This version of your project dropped calculation time
from 5½ min down to 8 sec !!!
Redirected to be useful w/o bug.

Jean

Forum Radovan_1.sm (37.48 KiB)已下载 1509 次。
#2 已发布: 2016/10/19 04:27:26
Radovan Omorjan

Radovan Omorjan

325个赞和1个踩,共2,052个帖子。

组: Moderator

Thank you Jean for this example,

I like it and will use it for my students.

Of course, this is not working in the recent SMath due to the mentioned bug about summation.
Never mind, it will work just with the few iteration variables changed until, Andrey fix the problem.

Best Regards,
Radovan
When Sisyphus climbed to the top of a hill, they said: "Wrong boulder!"
#3 已发布: 2016/10/19 10:59:57
Jean Giraud

Jean Giraud

983个赞,共6,866个帖子。

组: User

In Mathcad as well, we have encountered instances
that won't chain the calculations. The cure is to
isolate via spooling to file or redefining.

Worth trying this option.

Jean

Forum Radovan TryThis.sm (37.96 KiB)已下载 1587 次。
1位用户喜欢此帖子
Radovan Omorjan 2016/10/19 11:06:00
#4 已发布: 2016/10/19 11:08:02
Radovan Omorjan

Radovan Omorjan

325个赞和1个踩,共2,052个帖子。

组: Moderator

Nice, thank you

Thank you for your continuing contribution

Best wishes Jean,
Radovan
When Sisyphus climbed to the top of a hill, they said: "Wrong boulder!"
#5 已发布: 2016/10/19 11:14:40
Andrey Ivashov

Andrey Ivashov

2,344个赞,共4,057个帖子。

组: Super Administrator

Wrote

Of course, this is not working in the recent SMath due to the mentioned bug about summation.


Fix is easy - use line (BTW takes less then 2 seconds to calculate in latest SS).

Forum Radovan_1_mod.sm (36.29 KiB)已下载 1492 次。
#6 已发布: 2016/10/19 12:49:45
Radovan Omorjan

Radovan Omorjan

325个赞和1个踩,共2,052个帖子。

组: Moderator

Wrote

Wrote

Of course, this is not working in the recent SMath due to the mentioned bug about summation.


Fix is easy - use line (BTW takes less then 2 seconds to calculate in latest SS).

Forum Radovan_1_mod.sm (36.29 KiB)已下载 1492 次。



Hello Andrey,

Sorry, but I really do not understand what you wanted to say,and I am not sure if we understand each other. Let's get back to the post by Davide:

Summation variable problem

It is obvious that the problem exists regarding summation variable. Somehow, the summation variable gets its value from the global variables. I think it might be a source of confusion and prune to errors if the situation remains this way.

Regards,
Radovan
When Sisyphus climbed to the top of a hill, they said: "Wrong boulder!"
#7 已发布: 2016/10/19 13:00:12
Andrey Ivashov

Andrey Ivashov

2,344个赞,共4,057个帖子。

组: Super Administrator

Yes, this issue is well known and last several Months I was not able to fix it (tried many times).
That means I do not think this behavior will be corrected in nearest days/weeks.

But, in my comment you can find simple workaround which can be used - this is actually what I wanted to say.
1位用户喜欢此帖子
Radovan Omorjan 2016/10/19 13:02:00
#8 已发布: 2016/10/19 13:10:53
Radovan Omorjan

Radovan Omorjan

325个赞和1个踩,共2,052个帖子。

组: Moderator

I understand now, thank you.

I bumped into it quite some time ago, and did not understand it. I found the workaround, but was confused because my previous worksheets refused to work. I had to admit that I was I bit frustrated about it. At least, I hope I will remember what is going on regarding this problem. Especially, using line can do the thing, but it is a bit unusual sometimes.

I hope and wish you all the best in resolving this sooner or later .

Best Regards,
Radovan
When Sisyphus climbed to the top of a hill, they said: "Wrong boulder!"
#9 已发布: 2016/10/20 01:54:06
Jean Giraud

Jean Giraud

983个赞,共6,866个帖子。

组: User

True: it's getting confusing. Read this document carefully and experiment.
It finally points the code error. Whether an Smath code error or an error
from the borrowed summation code from foreign source. I had a bit of trouble
to make the 'ksmooth' work, lot more involved than this simple summation.

The Smath 5346 may reflect in other build past 5346.

Davide explanation is interesting as well as Andrey 'line'
but none answers/addreses the problem. As you will open the
attached document, if it does not work: then there is a bug
because 5346 works perfect saved in the "red star" construction.

Cheers, Jean

Maths Compute Pitfall.sm (47.84 KiB)已下载 1521 次。
1位用户喜欢此帖子
Radovan Omorjan 2016/10/20 04:07:00
#10 已发布: 2016/10/20 04:13:09
Radovan Omorjan

Radovan Omorjan

325个赞和1个踩,共2,052个帖子。

组: Moderator

Thank you Jean,

It also works well, and thank you for the discussion. By the way, I think that many times has been mentioned here the workaround of using line environment. I just often forget that, sorry. You mentioned "immensely long" expansions. That is one of the problem (trap) we might get into without to much thinking and using some many terms summation etc. Here, line might help to avoid such things.

Regards,
Radovan
When Sisyphus climbed to the top of a hill, they said: "Wrong boulder!"
#11 已发布: 2016/10/20 12:33:33
Jean Giraud

Jean Giraud

983个赞,共6,866个帖子。

组: User

Wrote

Thank you Jean,

It also works well, and thank you for the discussion. By the way, I think that many times has been mentioned here the workaround of using line environment. I just often forget that, sorry. You mentioned "immensely long" expansions. That is one of the problem (trap) we might get into without to much thinking and using some many terms summation etc. Here, line might help to avoid such things.

Regards,
Radovan



You have several points:

1. What you call "line" is what I call "compactum"
2. I have compacted two more functions. This has reduced the original
36 'Sol' 8 sec down to 6 sec 38 'Sol'
3. Whether 'line style' or simple does not reduce the expansion.
4. Horner expansion is not my cup of tea ! I need be educated on that.
5. the "optional" should work for you and other Smath build. Does it ?

BTW: such a nice piece of work would go well with a bit of abstract/source.

Cheers, Jean

Solve INVERSE Recursive.sm (62.18 KiB)已下载 1588 次。

1位用户喜欢此帖子
Radovan Omorjan 2016/10/20 14:52:00
  • 有新帖
  • 无新帖