PlanetMath (more info)
 Math for the people, by the people. Sponsor PlanetMath
Encyclopedia | Requests | Forums | Docs | Wiki | Random | RSS  
Login
create new user
name:
pass:
forget your password?
Main Menu
Owner confidence rating: High Entry average rating: No information on entry rating
hyperplane arrangement (Definition)

Let $V$ be a vector space over a field $\mathbb{K}$ A hyperplane arrangment in $V$ is a family $\mathcal{A}=\{\mathcal{H}_i\}_{i\in I}$ of affine hyperplanes in $V$ If all of the hyperplanes pass through $0$ $\mathcal{A}$ is called central; otherwise, it is affine. More generally, a subspace arrangement is a family of affine subspaces of $V$ The same distinction between central and affine subspace arrangement holds.

Example 1   Let $V=\mathbb{K}^n$ Then the family $$ \mathbb{K}P^n=\{S\subset V\mid\dim_{\mathbb{K}}(S)=1\} $$ of $1$ dimensional subspaces of $V$ is a central subspace arrangement, the projective space of dimension $n$ over $\mathbb{K}$

Instead of considering all lines through a vector space, we could consider all $k$ dimensional subspaces of the space.

Example 2   Again let $V=\mathbb{K}^n$ and suppose $0\le k\le n$ Then the family $$ \Gr(V,k)=\{S\subset V\mid\dim_{\mathbb{K}}(S)=k\} $$ of $k$ dimensional subspaces of $V$ is a central subspace arrangement, the Grassmannian. Observe that $\mathbb{K}P^n=\Gr(\mathbb{K}^n,1)$

If $V$ is a topological vector space and $\mathcal{A}$ is a hyperplane arrangement, then it makes sense to ask for the fundamental group of the complement $V\setminus\bigcup_{\mathcal{H}\in\mathcal{A}}\mathcal{H}$

Example 3   If $\mathcal{A}$ is a finite hyperplane arrangement over $V=\mathcal{R}^n$ then the arrangement partitions $V$ into a finite number of contractible cells. By selecting a point in each cell and taking the convex hull of the result, we obtain a polytope combinatorially equivalent to the zonotope dual to the arrangement. Since the question of the fundamental group here is not interesting, we could also use the embedding $\mathbb{R}^n\hookrightarrow\mathbb{C}^n$ to complexify $\mathcal{A}$ In this case the complement $\mathbb{C}^n\setminus\bigcup_{\mathcal{H}\in\mathcal{A}}\mathcal{H}$ usually has nontrivial fundamental group.

Bibliography

1
Klain, D. A., and G.-C. Rota, , Introduction to geometric probability, Cambridge University Press, 1997.
2
Orlik, P., and H. Terao, Arrangements of hyperplanes, Springer-Verlag, 1992.




"hyperplane arrangement" is owned by mps.
(view preamble | get metadata)

View style:

See Also: zonotope

Other names:  subspace arrangement, central arrangement
Also defines:  Grassmannian
Keywords:  arrangement, partition, fundamental group

Attachments:
intersection semilattice of a subspace arrangement (Definition) by CWoo
Log in to rate this entry.
(view current ratings)

Cross-references: embedding, zonotope, polytope, convex hull, point, cells, contractible, number, finite, complement, fundamental group, topological vector space, lines, projective space, subspaces, affine subspaces, pass through, affine hyperplanes, hyperplane, field, vector space
There are 4 references to this entry.

This is version 3 of hyperplane arrangement, born on 2006-03-23, modified 2006-03-23.
Object id is 7760, canonical name is HyperplaneArrangement.
Accessed 4014 times total.

Classification:
AMS MSC52C35 (Convex and discrete geometry :: Discrete geometry :: Arrangements of points, flats, hyperplanes)
 52B99 (Convex and discrete geometry :: Polytopes and polyhedra :: Miscellaneous)

Pending Errata and Addenda
None.
Discussion
Style: Expand: Order:
forum policy
Metric on the Grassmannian? by Schneemann on 2007-05-09 14:12:46
Hi.
Is it possible to define a metric on the Grassmannian, or a hyperplane arrangement? I am trying to figure that out for myself, but it really would help to find some reliable resources.

A metric would be a very simple way to obtain a topology.

Another question:
Consider a continuous mapping f_0 from a domain in RR^n to the Grassmannian Gr(n,k_0), and f_1 another continuous mapping from the same domain to the Grassmannian Gr(n,k_1).
The two maps could be a tangent field of the domain, and an arbitrary field of k_1-dimensional vector spaces, for instance.
Now we generate a new mapping by building the intersection (or the sum, alternatively) of f_0 with f_1. What are the necessary and sufficient conditions to make f_0 \cap f_1 continuous again? I would guess that f_0 \cap f_1 is continous in every subdomain where its dimension is constant - but again, it would help to have some reliable reference.

Thx.
[ reply | up ]

Interact
post | correct | update request | add derivation | add example | add (any)