Hi. I can't reproduce this in SMath.



File not found.File not found.

Best regards.
Alvaro.

Wrote

I can't reproduce this in SMath.

1/1 in Smath 6179, proof: the snippet.

Cheers Alvaro



Integrate with ode solver.sm (29.07 KiB) downloaded 676 time(s).

Wrote

I can't reproduce this in SMath.

You may not be able to implement the Mathcad bloc Given/ODEsolve.
ODEsolve is based on linterp. That is the first point.
The other point is that in Mathcad "linterp(x)" is a scalar function.
In Smath: ainterp, cinterp, linterp are simply interpolation fncts
They are not scalar functions. How does it reflects in ODEsolve ?
In Smath, ln(x), exp(x), ... are scalar from source code.
I tried to make cinterp scalar like in Mathcad, not to avail.

Please, persist on such a beautiful idea.

Wrote

I tried to make cinterp scalar like in Mathcad, not to avail.

This document was essentially designed to plot scalar the derivative on canvas,
as well as get a discrete export. Also included: the scalar quadratic.

Cheers ... Jean

Spline Deriv Infinitesimal.sm (70.87 KiB) downloaded 707 time(s).

Wrote

You may not be able to implement the Mathcad bloc Given/ODEsolve.

Hi Jean. Actually not, just the solver, without the given.
I found how to call the odesolver, with D(x,y)=[f(x)].
This method is much more faster than the direct integration.
But remains the ln(0) error, which seems to be a bug in the odesolvers, because zero isn't in the domain range.



Integrate with ode solver.sm (10.17 KiB) downloaded 639 time(s).

Best regards.
Alvaro.

The zero point is used to determine the dimension of the vector D(t,x) at runtime (D(0, x0)). The size of the system must be determined for the further operation of the numerical algorithm. It seems that it would be correct to substitute the beginning of the interval. I will try to change this. Also, if possible, I will support the one-dimensional case.

Thanks Viacheslav. It's amazing the effort that you and the others plugins developers put in enhance SMath, with very spectacular results.
Best regards.
Alvaro.

Wrote

Hi Jean. Actually not, just the solver, without the given.
I found how to call the odesolver, with D(x,y)=[f(x)].
This method is much more faster than the direct integration.
But remains the ln(0) error, which seems to be a bug in the odesolvers, because zero isn't in the domain range.

Buenos dias Alvaro !

ln(0) was introduced in Mathcad 11 soon after release via update [from recollection]
There is more in the attached, specially wrt parametric D(,,,).
Jean



ODE 0_Model General.sm (29.47 KiB) downloaded 685 time(s).

...

... refreshed version wrt explore all solvers.

ODE 0_Model General.sm (38.44 KiB) downloaded 685 time(s).