2
$\begingroup$

If we write a manifold or CW-complex $X$ as a subset of $\mathbb{R}^n$, in expression of coordinates, for example, \begin{multline} F(S^2,k+1)=\{(x_1,x_2,x_3,\cdots, x_{3k+1},x_{3k+2},x_{3k+3})\in\mathbb{R}^{3k+3}\\ \mid x_1^2+x_2^2+x_3^2=1,\cdots, x_{3k+1}^2+x_{3k+2}^2+x_{3k+3}^2=1,\\ \text{ for }i\neq j, x_{3i+1}\neq x_{3j+1} \text{ or } x_{3i+2}\neq x_{3j+2} \text{ or }x_{3i+3}\neq x_{3j+3} \}, \end{multline} is there any computer software or programming that can give the cohomology algebra automatically?

Can the computer give a very complicated simplicial complexes to approximate the manifold and compute the cohomology algebra?

$\endgroup$
  • $\begingroup$ This is not explicit in the question, perhaps on purpose, but the very specific space $F(S^2,k+1)$ is also known as the configuration space of $k+1$ points on $S^2$. Googling around shows that there supposedly already is a way (Totaro 1996) to compute the rational cohomology algebra of this space using spectral sequence methods. I guess the thrust of the question is whether an automated approach to this computation (if one is possible) would be advantageous to this alternative. $\endgroup$ – Igor Khavkine Mar 11 '15 at 13:08
1
$\begingroup$

It surely depends on what expressions you want to allow, but what comes to my mind is that even the much more basic question whether a given set is empty or not

  • is NP-complete, so don't expect an effective algorithm,
  • is computable by Tarski's theorem.

Maybe a logician could say more.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.