1 Pages (5 items)
Symbolic vs Numeric Optimization - Messages
#1 Posted: 3/11/2011 9:26:57 PM
What is the advantage to Symbolic Optimization?
It seems to me some things don't work right if Symbolic is choosen, which is the default.
For example, I created 2 user functions. Then I created another user function that included the first 2 functions. The result was 0 although I knew it should be some value. When I changed Optimization to Numeric the correct result was calculated.
I've noted similar posts where changing optimization to Numeric solved the problem.
I vote for Numeric as default.
It seems to me some things don't work right if Symbolic is choosen, which is the default.
For example, I created 2 user functions. Then I created another user function that included the first 2 functions. The result was 0 although I knew it should be some value. When I changed Optimization to Numeric the correct result was calculated.
I've noted similar posts where changing optimization to Numeric solved the problem.
I vote for Numeric as default.
Ed
#2 Posted: 3/12/2011 4:35:24 AM
Hello Ed,
I am not an expert in these things but, as far as I know, SMath did not have the "Optimization" in the previous versions. The main engine was "Symbolic" and everything goes through it - and still does. I can not explain what that means exactly. Just could guess that SMath uses some kind of library for manipulating expressions. This might be Andreys product or someone elses - do not know. In the meantime some problems occur by pure numerical results and Andrey introduced eval() function and "Numerical" or "None" optimization in the subsequent versions. As I could understood, the engine remains the same but some "shorcuts" have been introduced in order to strugle with numerical results. There still exist a problem with "symbolic" and "numeric" things in SMath. There are some calculations you can not do in SMath which are conected with numerical results. I do not know how would Andrey solve all these things, but I hope he will find the way.
Ed, you mentioned that "Numeric" optimization might be by default. I agree and could not see any problem by that. A bit strange might be that you can not see the pure numerical result (a number) when you hoover the mouse over an expression or use the Dynamis assitance. It is still in "Symbolic" form - as a fraction etc.
Andrey can give us the most accurate and complete answer about this.
Regards,
Radovan
I am not an expert in these things but, as far as I know, SMath did not have the "Optimization" in the previous versions. The main engine was "Symbolic" and everything goes through it - and still does. I can not explain what that means exactly. Just could guess that SMath uses some kind of library for manipulating expressions. This might be Andreys product or someone elses - do not know. In the meantime some problems occur by pure numerical results and Andrey introduced eval() function and "Numerical" or "None" optimization in the subsequent versions. As I could understood, the engine remains the same but some "shorcuts" have been introduced in order to strugle with numerical results. There still exist a problem with "symbolic" and "numeric" things in SMath. There are some calculations you can not do in SMath which are conected with numerical results. I do not know how would Andrey solve all these things, but I hope he will find the way.
Ed, you mentioned that "Numeric" optimization might be by default. I agree and could not see any problem by that. A bit strange might be that you can not see the pure numerical result (a number) when you hoover the mouse over an expression or use the Dynamis assitance. It is still in "Symbolic" form - as a fraction etc.
Andrey can give us the most accurate and complete answer about this.
Regards,
Radovan
When Sisyphus climbed to the top of a hill, they said: "Wrong boulder!"
#3 Posted: 3/12/2011 12:54:17 PM
This is another common occurance with SMath Symbolic Optimization.

If I switch to Numeric if works fine.

If I switch to Numeric if works fine.
Ed
#4 Posted: 3/12/2011 7:04:01 PM
Hello Ed,
As I remember, after some similar problem reported by a SMath user some time ago Andrey introduced "Optimization". There were other situations reported when the same error message appear, and the numerical optimization gave the result.
Regards,
Radovan
As I remember, after some similar problem reported by a SMath user some time ago Andrey introduced "Optimization". There were other situations reported when the same error message appear, and the numerical optimization gave the result.
Regards,
Radovan
When Sisyphus climbed to the top of a hill, they said: "Wrong boulder!"
1 Pages (5 items)
-
New Posts
-
No New Posts