## How would you define a hyper-cylinder? [closed]

I'd like to describe a "bottleneck" in n-dimensional topological vector space $\mathbb R^n$. For that I should probably try to define an "n-dimensional" cylinder.

I'd like to define something similar to a "hollow" cylinder but in say $\mathbb R ^n$ for n >= 2. I would probably define it as something homeomorphic to $S^{n-2} \times L$ where $L$ is a line crossing 0 (for a cylinder with finite volume I'd then assume this line to have two endpoints). Is there a name for this in Geometry or a reference where such a thing is discussed? Would this be the right way of defining a cylinder in n-dimensions? Any other suggestions?

-
This web site is for current research questions in mathematics. Your question is too elementary. – Ben McKay Dec 1 2011 at 10:42
What would such a cylinder be used for? In which context? – Qfwfq Dec 1 2011 at 10:59
I'd use it for pathplanning problem in an n-dimensional manifold. I would like to define a situation why even probabilistic complete pathplanning algorithms or analytic path planning algorithms would have difficulty (like the "bottleneck" phenomenon). This is a "current research question" in real algebraic geometry, but I'd like to start with something palatable in this site because I do not expect everyone to be specialized in this and still I want a broad discussion and perspectives from people from other fields. – Jose Capco Dec 1 2011 at 11:32
If you want to use it to represent an "uncertain" curve, then it's simply a tubular neighborhood of a curve or maybe just its boundary. – Deane Yang Dec 2 2011 at 4:53

## closed as off topic by Willie Wong, Ryan Budney, Qiaochu Yuan, Andy Putman, Deane YangDec 2 2011 at 4:52

I believe an abstract cylinder is any manifold having the form $M \times I$, where $I \subseteq \mathbb{R}$ is an interval. This is called a 'cylinder over $M$'. In your case, you'd want $M$ to have codimension 2 in the ambient space. Your definition is compatible with this.