Reply
 
LinkBack Thread Tools Search this Thread Display Modes
  #1   Report Post  
Amitava
 
Posts: n/a
Default solver can not solve polynomial eq. where one variable is five di.

I tried to create equations from the set of data by solver
x= 90021 85248 78314 69768 59508 47975
y = 20.09 19.92 18.70 16.95 16.13 15.67

in the form of y= a + bx +cx2 + dx3;
where x2= x square ,x3 = x cube.

but could not do so.
the solver does fine when the value of x is small .
Your help will be hghly appreciated in this regard.

Thanks
  #2   Report Post  
tjtjjtjt
 
Posts: n/a
Default

You might want to check out solver.com to find out exactly what the
limitations and weak points for the buitl-in Solver are.

tj

"Amitava" wrote:

I tried to create equations from the set of data by solver
x= 90021 85248 78314 69768 59508 47975
y = 20.09 19.92 18.70 16.95 16.13 15.67

in the form of y= a + bx +cx2 + dx3;
where x2= x square ,x3 = x cube.

but could not do so.
the solver does fine when the value of x is small .
Your help will be hghly appreciated in this regard.

Thanks

  #3   Report Post  
Harlan Grove
 
Posts: n/a
Default

"Amitava" wrote...
I tried to create equations from the set of data by solver
x= 90021 85248 78314 69768 59508 47975
y = 20.09 19.92 18.70 16.95 16.13 15.67

in the form of y= a + bx +cx2 + dx3;
where x2= x square ,x3 = x cube.

but could not do so.
the solver does fine when the value of x is small .
Your help will be hghly appreciated in this regard.


In Excel you're limited to 15 decimal digit precision most of the time. A
few intermediate calculations may have more if they can stay in FPU
registers. However, standard minimizing sum squared error fitting would at
some point involve squaring the x^3 values, at which point you've blown way
past machine precision.

That said, plot these X-Y values in an X-Y chart, add a polynomial trend
line or order 3, and display the trend line equation. It'll show the a, b, c
and d coefficients you seek.


  #4   Report Post  
Bernard Liengme
 
Posts: n/a
Default

You can use LINEST to find equation of best fit.
See www.stfx.ca/people/bliengme/ExcelTips

--
Bernard Liengme
www.stfx.ca/people/bliengme
remove CAPS in email address


"Amitava" wrote in message
...
I tried to create equations from the set of data by solver
x= 90021 85248 78314 69768 59508 47975
y = 20.09 19.92 18.70 16.95 16.13 15.67

in the form of y= a + bx +cx2 + dx3;
where x2= x square ,x3 = x cube.

but could not do so.
the solver does fine when the value of x is small .
Your help will be hghly appreciated in this regard.

Thanks



  #5   Report Post  
Tushar Mehta
 
Posts: n/a
Default

I imagine you could tweak Solver's parameters and/or scale your own
values to get Solver to work, but, in this specific instance, you
should consider using LINEST as Bernard suggested.

If you did, you would discover that none of the coefficients for an
order 3 fit are statistically significant at the 5% level. Only the
constant term is significant for a quadratic fit whereas both terms are
for a linear fit. To learn more about how to test for statistical
significance check the LINEST help Example 5.

--
Regards,

Tushar Mehta
www.tushar-mehta.com
Excel, PowerPoint, and VBA add-ins, tutorials
Custom MS Office productivity solutions

In article ,
says...
I tried to create equations from the set of data by solver
x= 90021 85248 78314 69768 59508 47975
y = 20.09 19.92 18.70 16.95 16.13 15.67

in the form of y= a + bx +cx2 + dx3;
where x2= x square ,x3 = x cube.

but could not do so.
the solver does fine when the value of x is small .
Your help will be hghly appreciated in this regard.

Thanks

Reply
Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules

Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On



All times are GMT +1. The time now is 06:21 PM.

Powered by vBulletin® Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004-2024 ExcelBanter.
The comments are property of their posters.
 

About Us

"It's about Microsoft Excel"