5
$\begingroup$

Let $Y$ be a $1$-reduced finite simplicial set, $\Sigma^n$ be a simplicial model of the $n$-sphere and $f: \Sigma^n\to Y$. We further know that $f$ is nullhomotopic (this can be algorithmically verified via obstruction theory and Postnikov towers such as in this paper).

Our goal is to construct an explicit simplicial extension $F: B^{n+1}\to Y$ of $f$ such that $B^{n+1}$ is a triangulation of the $(n+1)$-disc and $\partial B^{n+1}$ is a simplicial subdivision of $\Sigma^n$.

Such algorithm clearly exists, as one could add one vertex to $\Sigma^n$ to create a simplicial model of $B^{n+1}$, then do iterative barycentric subdivisions and "try" all simplicial maps to $Y$ and eventually succeed.

Is there a more feasible algorithm with, let's say, polynomial complexity?

If we replace $Y$ via its Postnikov tower $P_{n+1}$ and replace $f$ by $(\varphi f)$ where $\varphi: Y\to P_{n+1}$ is $n$-connected, then we can construct an extension $B^{n+1}\to P_{n+1}$ explicitely, see, for example, this paper. However, we don't see how to find a nullhomotopy/extension to $Y$, which in general is not Kan.

$\endgroup$
9
  • $\begingroup$ Well, the first problem is that the volume of the filling ball may be exponential in the volume of the sphere. $\endgroup$
    – HJRW
    Commented Jan 28, 2017 at 12:31
  • $\begingroup$ @HJRW Thanks; can you please elaborate on that? Actually, I don't even see how to create an "exponential" filling (or how to prove it exists). $\endgroup$ Commented Jan 28, 2017 at 20:41
  • $\begingroup$ The standard example occurs in the presentation complex of the $(1,2)$-Bausmlag--Solitar group $\langle a,b\mid bab^{-1}=a^2\rangle$. The loop corresponding to the word $a^{-1}b^kab^{-k}ab^{k}a^{-1}b^{-1}$ is filled by a disk with area of the order of $2^k$. (And worse, in any group with unsolvable word problem, the filling area is a non-recursive function of the length of the loop.) See Proposition 3.2.2 of Bridson's article 'The geometry of the word problem'... $\endgroup$
    – HJRW
    Commented Jan 28, 2017 at 21:22
  • $\begingroup$ ... The article also contains lots of references. Much less is known about 'higher dimensional' filling functions, but I'm fairly certain they can be superpolynomial. people.maths.ox.ac.uk/bridson/papers/bfs $\endgroup$
    – HJRW
    Commented Jan 28, 2017 at 21:23
  • 1
    $\begingroup$ On reflection, the literature I'm thinking of is about aspherical spaces. Probably the suspension of the presentation complex of the Bamslag--Solitar group gives an example with exponential 2nd filling function. And similarly, the suspension for a group with worse Dehn function should give a larger 2nd filling function. $\endgroup$
    – HJRW
    Commented Jan 29, 2017 at 10:29

0

You must log in to answer this question.

Browse other questions tagged .