Computing "near isometries"

29 Views Asked by At

Given a finite dimensional inner product space V, and a collection of vectors $\mathcal{C}$ in V, are there any algorithms for finding nontrivial isometries of V which map $\mathcal{C} \to \mathcal{C}$ (or confirming that non exist)? I am actually interested in this in a case where I would only need my isometry to "nearly" preserve $\mathcal{C}$. There are various ways you could make the phrase "nearly preserve $\mathcal{C}$" precise, and I agnostic about how you formalize this.