Different types of solvers

[Ramax]

Hi guys,

I'm just getting started with Mathcad. Currently I'm trying to understand the different kinds of solvers. It seems to me there are three "basic categories":

1. "Single command Solvers", e. g. root, lsolve etc. where you just pick your appropriate solver and launch

2. "Solve blocks", which allow for separate guess values and boundary constraints, but seem to support only find, minerr, minimize, maximize and odesolve

3. The symbolic "solve" command, which seems to be for solving equations symbolically (but apparently also produces numerical solutions).

I don't quite understand what the latter two are for. What is their benefit? Aren't there plenty of solvers for everything (save maybe symbolic solutions) in the first "category"? So when should I use what? 🤔 *confused*

Any help would be much appreciated.

Thanx.

[Alex_Cazacu]

Hello,

Have you looked in the available trainings:

https://learningconnector.ptc.com/products/mathcad

You may find the answers there.

Thanks,
Alex

[Ramax]

Hi Alex,

thanks for your response. Of course I can do that. The problem is that it's rather tedious to dig through everything to find a specific answer. Also from my experience, tutorials mostly just introduce features and do not compare them among each other. But if there is no specific answer at hand, that is probably the way to go.

[Alex_Cazacu]

Hi Ramax,

Here is what I could find out:

In "Solve blocks" the user can define relations and insert constraints - limits of calculation and then, in the solver part the equation(s) will be solved depending on the "Single command Solvers", so the main difference between these two are constrains.

The Symbolic Solve is used if an equation is solved symbolically, it will show the roots of the equation. If some of the variables are already defined previously in the worksheet, then it will be taken into consideration and the roots will been displayed accordingly, depending on those too.

The user can choose from a variety of ways to solve depending on what he finds to be more accessible, more easy!

I hope this helps!

Thanks,

Alex

[Ramax]

Hi Alex,

tanks again. That definetly does help. So it seems like all three in principle are able to do the same (save that "solve blocks" allow for relations and constraints and "solve" can solve symbolically). Can you say something about efficiency? I imagine there shouldn't be much of a difference between "single commands" and "solve bocks" (at least as long as you use the same solver). How about using symbolic "solve" for a numerical solution? Would that be less efficient or is there basically no difference?

[Alex_Cazacu · Accepted solution]

Hello,

As I understood, they all do the same thing, but they can be used in different situations. I'd say that the most efficient is the one you are most comfortable with, unless you need more functionality this one can't handle.

Thanks,
Alex

[Ramax]

Hi Alex,

thanks a lot. That gives me a standard to work by.