Grassmannian

The Grassmannian, denoted Gr(k,m), is the collection of every k‑dimensional flat inside an m‑dimensional vector space. Rather than listing each subspace individually, mathematicians treat this whole set as a single geometric object that itself has shape: it can be given a smooth structure, so nearby points in the Grassmannian correspond to subspaces that differ only slightly.

Because it packages together all possible linear directions of a given size, the Grassmannian shows up whenever one needs to compare or optimise over subspaces. It underlies many constructions in algebraic geometry where one studies families of lines, planes, and higher‑dimensional flats; in representation theory as a natural habitat for certain group actions; and in engineering fields such as signal processing and computer vision, where problems often reduce to choosing the best subspace for data fitting or dimensionality reduction.

In physics the Grassmannian appears when describing collections of quantum states that share a fixed number of excitations, and in modern optimization it provides the backdrop for algorithms that move along the space of low‑rank matrices. In each of these settings the idea is the same: treat the set of subspaces not as a loose list but as a coherent geometric entity whose structure can be studied and exploited.