I'm trying to find a numerical instance (I don't need arbitrarily high precision) of solution to 7 relatively complicated inequalities (plus the region specification) involving 8 variables.

The problem is that the kernel quits itself while evaluating the FindInstance. I have tried to fix a finite WorkingPrecision, to no avail.

I understand that this is probably too little information to even suggest something, but unfortunately I cannot produce a MWE -- the notebook and expressions are too complicated and refer to other files.

Is there any general pointer you could intuitively give me? I was thinking about researching for an ideal Method, but don't even know where to begin.

I appreciate any insights.

  • 1
    $\begingroup$ This post is relevant. $\endgroup$
    – yarchik
    Oct 15, 2021 at 17:05
  • 1
    $\begingroup$ Is there any chance that you could, perhaps by trial and error, pick a subset of your inequalities that constrains where your final solution might be, find the solution to that relaxed problem, and then use that solution to help narrow down the space to look for your full solution? $\endgroup$
    – Bill
    Oct 15, 2021 at 17:17
  • 1
    $\begingroup$ Try, if you succed with cd = CylindricalDecomposition[inequalities, variables] . Then FindInstance[cd, variables] is an easy job. $\endgroup$
    – Akku14
    Oct 15, 2021 at 18:25
  • $\begingroup$ Thank you all for the suggestions -- I'll try them as soon as possible. $\endgroup$
    – GaloisFan
    Oct 15, 2021 at 19:47


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.