I have a set S of high dimensional points in Euclidean space, with convex hull H (not known); and some target point T in that space not in or on H.
Rather than worry about calculating both H and the distance to it explicitly, is there an efficient way to calculate just the distance from T to the closest point on H? I was thinking something involving binary space partitioning with spitting planes, but can't quite figure out how to formulate the approach (esp. which planes to choose).