Graph-based Scheduling Algorithms for Mobile Workstations

Abstract

The success of Kiva robots, which move warehouse shelves towards human workers to speed up delivery times in distribution centers, inspires us to apply the same idea to service robots and propose the concept of mobile workstations, which is basically a combination of production machinery and a mobile platform with additional capabilities of fetching and delivering objects. Mobile workstations, which process materials while moving towards to their destinations, can substantially shorten the time to deliver a product by paralleling the movement of materials and the processing of materials. We define a general model of mobile workstations and devise two graph-based temporal planning algorithms to control a mobile workstation to handle job requests over time. Our experimental results show that the local search algorithm can greatly decrease the running time of the graphbased planning algorithm for handling a large number of jobs while the temporal plans it generates have a makespan that is close to optimal in small problems.