Dynamic programming solution for a class of pursuit evasionproblems: the herding problem

摘要

A herding dog and sheep problem is studied where the agentn“dog” is considered the control action for moving the agentn“sheep” to a fixed location using the dynamics of theirninteraction. The problem is solved for the deterministic case usingndynamic programming. Proofs are provided for the correctness of thenalgorithms. The algorithm is analyzed for its complexity. A softwarenpackage developed for experimentation is described