Analysis of solutions to the time-optimal planning and execution problem

This paper analyses solutions to the time-optimal planning and execution (TOPE) problem, in which the aim is to minimise the total time required for an agent to achieve its objectives. The TOPE process provides a means of adjusting system parameters in real-time to achieve this aim. Prior work by the authors showed that agent-based planning systems employing the TOPE process can yield better performance than existing techniques, provided that a key estimation step can be run sufficiently fast and accurately. This paper describes several real-time implementations of this estimation step. A Monte-Carlo analysis compares the performance of TOPE systems using these implementations against existing state-of-the-art planning techniques. It is shown that the average case performance of the TOPE systems is significantly better than the existing methods. Since the TOPE process can be added to an existing system without modifying the internal processes, these results suggest that similar performance improvement may be obtained in a multitude of robotics applications.