In a metric space , a geodesic ray is an isometric embedding :Every finite segment therefore minimizes length between its endpoints. In Riemannian geometry, this is a globally minimizing unit-speed geodesic defined for all nonnegative times. Some authors use “ray” more loosely for any half-infinite geodesic, so the minimizing convention matters.
In Euclidean space, is a geodesic ray whenever . On a unit round sphere, a unit-speed great circle continues indefinitely as a geodesic, but it is not a ray: after time it returns to its starting point. In fact, a bounded metric space cannot contain a geodesic ray, since its points would have to lie arbitrarily far from .
New to topics? Read the docs here!