Given a finite set $S = \lbrace x: x \in R^n \rbrace $ of points, where the $i$-th point is associated with a weight $w_i \in [0,1]$ and all these points are spread across the surface of a n-sphere $S^n$. I perform a stereographic projection to $R^n$ and notice that there is also a distance metric (as shown below) that I can make use of: Metric on n-sphere in terms of stereographic projection coordinates

My problem is that I need a weighted-metric, whereby the weights associated with each point are also taken into account. So, for example, given any arbitrary point $p \in S^n$ and $d$ is the metric (unweighted) on the sphere, the nearest point $p^*$ to $p$ would minimize $d(p,p^*)$ and maximize $w_{p^*}$. In other words, the metric must favour points that are closer to $p$ on the sphere, and whose weights are the largest.

I am wondering if such a min-max objective function can be achieved in a metric. I was thinking of something simple like : $\alpha \cdot d(p,x) + (1-\alpha)\cdot w_x$ to start with. Are there better ways to do this on a Riemannian manifold like the sphere?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.