Base and prior solutions problem + solutions found do not satisfy equation I specified
• Hello,

I encountered two problems.

I specified that I am looking for a formula that satisfies
z = f1(erf(f2(x, y, log(x)))) + f3(erf(f4(x, y, log(x)))) + 0.5

I seeded this run with a prior solution that I pasted into the "base and prior solutions" field.  Here it is:
f0 = 0.5 + 0.0125240211293536*erf(1.34339047092964*log(x) - 1.26777649668828 - 2.12394504369926*y^2) + 0.487475978870646*erf((0.974158485628697*log(x) - 0.842297540173634 - 0.111666033842753*exp(y))/y)

When I tried running Eureqa, I immediately got an error message:

"Unable to interpret the prior solution ...

based on the current variable and Target Expression"

Here is what the log says

" Couldn't seed solution: 0.5 + 0.0125240211293536*erf(1.34339047092964*log(x) - 1.26777649668828 - 2.12394504369926*y^2) + 0.487475978870646*erf((0.974158485628697*log(x) - 0.842297540173634 - 0.111666033842753*exp(y))/y), 0, 0, 0, 0"

I believe that this solution satisfies z = f1(erf(f2(x, y, log(x)))) + f3(erf(f4(x, y, log(x)))) + 0.5

Would it be possible to identify what causes this problem?

I had encountered this problem before.  I would run Eureqa, and obtain a good solution.  Then I would rename the Eureqa project, paste that good solution into "base and prior solutions" field, not change any other parameter or setting, and try running Eureqa - I would then get the same error message.  I am not sure why I am getting an error message if I am plugging a solution obtained using the same set of variables and the same target expression.

My second problem is that the solutions returned by Eureqa when I specified z = f1(erf(f2(x, y, log(x)))) + f3(erf(f4(x, y, log(x)))) + 0.5 as my target expression, do not seem to fit the target expression.  Here are some of these solutions that I am copying from the Results tab:
"z = 0.479529121823671 + 0.520777354748168*erf(log(x) - 1.06406080128452)"
"z = erf(0.0306987792680305*x^2 + 0.159330745448271*x*erf((0.211474704895106*x^2*y^2 - 0.000398593888560477)^0.813694266913868) - 0.0335879197729986*y*x^2) - 0.00782916253708832"

Am I doing something wrong?  I tried to attach my .fxp file to this message, but I got an error message saying that files of this type are not allowed to be uploaded.

Thank you,

Bill

 Tweet
• Hi Bill,

Sorry for the delay - I was sourcing more info. Can you try putting parentheses around the pre-computed function inputs?

For example, the following:

z = f1(erf(f2(x, y, log(x)))) + f3(erf(f4(x, y, log(x)))) + 0.5

should be:

z = f1(erf(f2(x, y, (log(x)) ))) + f3(erf(f4(x, y, (log(x)) ))) + 0.5

because it is a determined input that gets pre-computed into a new variable with the same name but with parentheses.

Let me know if that helps.

Best,

Jess

#### Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!