1-AUG-2019 SolveSys - Equation Library and Nonlinear Systems Solver

[PLX]

I'll see if I can dig up the exact system that caused issues - I was a system of 7 or 8 equations with a couple being linear and the rest quadratic. It was usually fine, but sometimes would chug along for a couple minutes and then return complex values for all of the variables. If I solved the equations in blocks, I'd get it to converge after having the complex result. It seemed overwhelmed with the whole system. The guess values were "realistic" given the physical expectation of the system.

[akmon]

Please, can you check if this app works fine with latest 13333 firmware version? Mine is no able to start solving. Thanks.

[Han]

(02-03-2018 11:58 PM)akmon Wrote:  Please, can you check if this app works fine with latest 13333 firmware version? Mine is no able to start solving. Thanks.

http://www.hpmuseum.org/forum/thread-10067.html

I will be publishing a newer version soon that has some feature updates and will also provide a workaround for this bug.

[Uhtred]

(02-06-2018 07:27 PM)Han Wrote:  

(02-03-2018 11:58 PM)akmon Wrote:  Please, can you check if this app works fine with latest 13333 firmware version? Mine is no able to start solving. Thanks.

Ok thanks, your programm is so usefull but I have the same problem.

http://www.hpmuseum.org/forum/thread-10067.html

I will be publishing a newer version soon that has some feature updates and will also provide a workaround for this bug.

[arturocb]

Hello everybody, I would like some help about an issue a friend has had.

She got her HP Prime a month ago and has recently added this Equation Library App to it in order to store and solve different systems of equations we need for our engineering classes. She has done everything as I told her to, since I've had it for a couple of months and have not had one problem with it. I have not seen any weird settings; the calculator has been updated and I reset it a couple of times.

The problem is I add variables, define the equations I want to solve and when I hit the key NUM to give values to the different parameters, I get a message that says "No equations defined". Then, all the equations disappear as if they had never been typed in. The variables are still there but the equations aren't. I had never seen this kind of error, and I have worked with a lot of different classmates that have had problems with the HP Prime, all of them solved sooner or later.

Also, programs do not execute correctly, there are no errors at all but it won't let us get to the input window.

Any tips you can give me will be appreciated, we have no clue what's going on and I can't believe such an expensive calculator can give so many problems. Thank you very much!

[Han]

1) Which firmware do you have? (Press the Help key, then select the Tree menu option, and then choose About HP Prime from the list-menu.

2) What system of equations were you trying to solve.

[sinan oz]

(03-21-2018 10:15 PM)arturocb Wrote:  Hello everybody, I would like some help about an issue a friend has had.

She got her HP Prime a month ago and has recently added this Equation Library App to it in order to store and solve different systems of equations we need for our engineering classes. She has done everything as I told her to, since I've had it for a couple of months and have not had one problem with it. I have not seen any weird settings; the calculator has been updated and I reset it a couple of times.

The problem is I add variables, define the equations I want to solve and when I hit the key NUM to give values to the different parameters, I get a message that says "No equations defined". Then, all the equations disappear as if they had never been typed in. The variables are still there but the equations aren't. I had never seen this kind of error, and I have worked with a lot of different classmates that have had problems with the HP Prime, all of them solved sooner or later.

Also, programs do not execute correctly, there are no errors at all but it won't let us get to the input window.

Any tips you can give me will be appreciated, we have no clue what's going on and I can't believe such an expensive calculator can give so many problems. Thank you very much!

I had faced with similar problem:I chose Digit Grouping: 123.456,789 , and it happened.
It is solved when I chose 123,456.789

[Han]

I think I know what the problem is, now. My input forms are created dynamically, and I specifically use the "," symbol for separation when I should be using the user-selected symbol for fraction mark.

[Andrea]

I'm new using the HP Prime. I'm studying Engineering and I'm having trouble using the calculator, specially solving system of ecuations.

I have downloaded your application, but still can't solve the system.
It will mean a lot for me if you can help me.
The ecuations are:
0.05000767076=19.93*10^(-3)*V+8.81*10^(-3)*W

-(V^2/2)−76.2824057829+T-(0.25/(LOG((1/2455.875)+(5.74/(159300^0.9*V^0.9))))^2)*470.8097*V^2=0

121.530094217+T-(0.25/(LOG((1/1632.625)+(5.74/(105900^0.9*W^0.9))))^2)*1652.502*W^2=0

I need to know the values of V, W and T.

My friends have the 50-G and use the solvesys and the haven't had any problem.

Thank you

[akmon]

(09-28-2018 10:48 AM)Andrea Wrote:  I'm new using the HP Prime. I'm studying Engineering and I'm having trouble using the calculator, specially solving system of ecuations.

I have downloaded your application, but still can't solve the system.
It will mean a lot for me if you can help me.
The ecuations are:
0.05000767076=19.93*10^(-3)*V+8.81*10^(-3)*W

-(V^2/2)−76.2824057829+T-(0.25/(LOG((1/2455.875)+(5.74/(159300^0.9*V^0.9))))^2)*470.8097*V^2=0

121.530094217+T-(0.25/(LOG((1/1632.625)+(5.74/(105900^0.9*W^0.9))))^2)*1652.502*W^2=0

I need to know the values of V, W and T.

My friends have the 50-G and use the solvesys and the haven't had any problem.

Thank you

I´ve inserted the equations, and initial values V=1, W=1 and T= 1 and I got:
V= 1,4841386
W= 2,3188181
T= 101,06092
[F]= 8.959E-10

Are these results what you expected?

[rrpalma]

Hello,

I'm trying to solve the following system using SolveSys:
(X/Y)^Z = 1/8
2^X/3^Y=4/Z^4
X*Y/Z=8/3

On the HP50g, I just use the equation writer to place all 3 equations on the stack, then create an array with them. I then create an array with [X Y Z] and another with my initial guesses [1 1 1]. I go to NUM.SLV-->MSLV and then get an array with 2, 4 and 3 as the answers.

However, when trying to solve this system on the Prime using SolveSys, I get the following message: Newton iterations terminated but ||F|| failed tolerance
The (incorrect) answers that I get are X=1.505 Y=4.984 Z=2.759

What am I doing wrong?

[Han]

(05-05-2019 08:10 PM)rrpalma Wrote:  Hello,

I'm trying to solve the following system using SolveSys:
(X/Y)^Z = 1/8
2^X/3^Y=4/Z^4
X*Y/Z=8/3

On the HP50g, I just use the equation writer to place all 3 equations on the stack, then create an array with them. I then create an array with [X Y Z] and another with my initial guesses [1 1 1]. I go to NUM.SLV-->MSLV and then get an array with 2, 4 and 3 as the answers.

However, when trying to solve this system on the Prime using SolveSys, I get the following message: Newton iterations terminated but ||F|| failed tolerance
The (incorrect) answers that I get are X=1.505 Y=4.984 Z=2.759

What am I doing wrong?

Probably nothing. The message simply means that the pre-set number of iterations were reached but that the solution was still not obtained. I have not tried the system myself or even looked at the mathematical reasons, but just an initial gut feeling based on your numerical values I would say the system is slow to converge numerically. Just re-run the solver again and see if the values get closer to the solution found on the HP50G.

EDIT: What I mean is to run the solver using the current values for X Y and Z as your new initial guess. You can also change the max number of iterations in the options. Lastly, you should be able to see the result of each step to at least see if the convergence is slow.

[rrpalma]

Thanks Han for your promptly replying to my post.

You were right! It took three runs to find a zero. I used the previously found on run (n-1) variable values as guesses for run (n).

However, when I increased the number of iterations from 100 to 1000, the program did not converge on the first run.

Thanks a lot for pointing me in the right direction. This is a great app.

[OGee]

Thanks Han,

I've had the same problem as others: 'no equations...' due the ',' vs '.' digit grouping setup.

Is there an easy fix for this, so I don't need to switch number format when using Equation Library?

Thanks
Ole

[Han]

(07-30-2019 12:43 PM)OGee Wrote:  Thanks Han,

I've had the same problem as others: 'no equations...' due the ',' vs '.' digit grouping setup.

Is there an easy fix for this, so I don't need to switch number format when using Equation Library?

Thanks
Ole

Funny you should ask. I had been working on this and will have a test version up hopefully by the end of this week.

[OGee]

Hi Han,

6th sense

Thanks, looking forward for this.

Regards
Ole

[Han]

Here's a new beta to test. It only contains the separated version, so you will need to separately download the SVD program. (Sorry if this is an inconvenience but it is just temporary.) Make sure you take the necessary steps to back up your libraries if you modified the default library, or created new ones. See the LibToNote program in the eqlib.pdf manual file in the Ver. 1.00 release. See first post for more information.

Version 1.10Beta :
  Equation Library.hpappdir.zip (Size: 73.92 KB / Downloads: 490)

[ymmot]

Hello... So One should first install version 1 and then install the 1.10 beta?

[Han]

(03-14-2020 07:28 AM)ymmot Wrote:  Hello... So One should first install version 1 and then install the 1.10 beta?

You can just install the beta; sorry for such a late response (didn't notice it 'til now).

[ymmot]

Hi Han,
Many thx for a great program.
Is it possible to solve overdetermined equations systems with your Solvesys (more equations than unknowns)?
What about softbutton interface...
And it would be nice to have a cmd-line version that you could call from within your own porograms,