2021/11/11 by Abhinav Dahiya, Nima Akbarzadeh, Dahiya, Abhinav +5 · 1 citation
Decision Sciences · Computer Science · Engineering · #Advanced Bandit Algorithms Research #Optimization and Search Problems #Smart Grid Energy Management
paper · pdf · doi:10.48550/arxiv.2111.06437
In this paper, we consider the problem of allocating human operators in a\nsystem with multiple semi-autonomous robots. Each robot is required to perform\nan independent sequence of tasks, subjected to a chance of failing and getting\nstuck in a fault state at every task. If and when required, a human operator\ncan assist or teleoperate a robot. Conventional MDP techniques used to solve\nsuch problems face scalability issues due to exponential growth of state and\naction spaces with the number of robots and operators. In this paper we derive\nconditions under which the operator allocation problem is indexable, enabling\nthe use of the Whittle index heuristic. The conditions can be easily checked to\nverify indexability, and we show that they hold for a wide range of problems of\ninterest. Our key insight is to leverage the structure of the value function of\nindividual robots, resulting in conditions that can be verified separately for\neach state of each robot. We apply these conditions to two types of transitions\ncommonly seen in remote robot supervision systems. Through numerical\nsimulations, we demonstrate the efficacy of Whittle index policy as a\nnear-optimal and scalable approach that outperforms existing scalable methods.\n