Definite Integral Issue with Mathcad 15

[mshaw-3]

Hi there,

Not sure if I have an incorrect implementation/setup or if it's some kind of bug...

If I perform the following in Mathcad 2001i, it works instantly:

If I perform the same in Mathcad 15 it will not solve and I eventually I get the message "Condition contains Otherwise in subexpression".

I have attached the files I have been using for your perusal.

[MikeArmstrong]

I also get an error message, looks like a bug to me.

[mshaw-3]

Thanks.

Wasn't sure if I had any settings at error within the worksheet. I don't think I do, I literally loaded Mathcad and entered the equation.

If it is a bug - do the PTC devs look in on these forums? Or would I need to submit to somewhere first?

Thankyou.

[MikeArmstrong]

Mike Shaw wrote:

Thanks.

If it is a bug - do the PTC devs look in on these forums? Or would I need to submit to somewhere first?

Thankyou.

I think you will have to raise a Customer Support call to the Technical Support department. See http://www.ptc.com/cs/doc/copen.htm for how to contact PTC technical support.

[mshaw-3]

Thankyou. It seems I have to upgrade my Account (I purchased a Student Version for Open University work), but I don't have a Customer Number...

[MikeArmstrong]

Unfortunately I do not have those privileges either. Maybe someone who doe will report it for you. VladimirN.

is usually very helpful in these situations.

[ValeryOchkov]

Mathcad 2001i has a Maple symbolic engine

Mathcad 13-15 - MuPad symbolic engine

But Maple is Maple - the best symbolic tool!

[mshaw-3]

Interesting, I didn't know that - thankyou.

[RichardJ]

The MuPad symbolic engine has always been very weak when it comes to integrals, and especially definite integrals. In this case it can solve the indefinite integral, but not the definite integral, which is not logical. This is not the first example I've seen where this is the case though.

At a minimum, the error message is clearly a bug.

[StuartBruff]

Richard Jackson wrote:

The MuPad symbolic engine has always been very weak when it comes to integrals, and especially definite integrals. In this case it can solve the indefinite integral, but not the definite integral, which is not logical. This is not the first example I've seen where this is the case though.

Interesting. That's definitely something to bear in mind.

At a minimum, the error message is clearly a bug.

Maybe, Maybe Not. At a Mathcad level, probably, but it's possibly telling us something the internal error that MuPad encountered.

Stuart

[Werner_E · Accepted solution]

I agree with Richard and others that integrals and especially definite integrals alway were a nightmare when the symbolics changed to MuPad. I still remember the many clearly wrong results of rather simple definite Integrals in Mathcad 14 and the some even in the first release of MC15.

This one here is really bad, especially given that its an easy one and consequently the indefinite interal is no problem, not even for Mupad.

That way we can create a workaround here:

The problem may arise because Mupad defaults to domain complex and that way probably has to make more distinction of cases as we may think of. Not sure about that, though.

If I tell Mathcad/MuPad that all variables are real and positive, I get a result but its frightening and takes quite long to calculate! I am not even sure its correct and equivalent to the above. Not sure whats going on here.

.......

[VladimirN]

I have the following result (as in MC2001i) by using "simplify" and "assume" symbolical operators in the MC15 F000 service release:

[RichardJ]

In MC15 030 I get something different:

If I add z to the list I get the desired result:

If I also add L to the list I get a big, ugly answer that's similar to, but not the same as, the one Werner shows.

So the trick seems to be to force the domain of all the variables in the kernal, but not the limits.

[RichardJ]

Major symbolic weirdness! I just did it again. If I have only Z and X in the list I get the incomplete solution. Then I copy and paste the expression and add z to the list, and I get the desired solution. BUT, if I then disable calculation for the first expression and enable it again, it also gives the desired solution!

[StuartBruff]

Richard Jackson wrote:

Major symbolic weirdness! I just did it again. If I have only Z and X in the list I get the incomplete solution. Then I copy and paste the expression and add z to the list, and I get the desired solution. BUT, if I then disable calculation for the first expression and enable it again, it also gives the desired solution!

Does that mean it's Friday?

[RichardJ]

. Good point. Maybe this doesn't happen except on certain days. With Mathcad symbolics almost anything is possible!

Actually, I think I have seen this kind of "memory" behavior before, but not in a way where I could reproduce it at will.