Assign orders to workstations in robotic sorting systems considering destination information

Abstract

Robotic sorting system is the newest implementation in intralogistics to replace costly manual workers. Such a system can be flexibly designed on a two-tier platform where robots derive on the top tier and sort parcels from workstations to drop-off points on spiral conveyors connected to roll containers at the lower tier. While enjoying higher throughput and lower operating cost than a traditional conveyor-based sorting system, such systems still face many challenging issues. One of those is the assignment of orders to workstations. We build a semi-open queueing network to estimate the system performance and design a heuristic to improve the assignment rule based on two benchmark rules. The analytical model is validated through simulation. Numerical results show that our heuristic procedure can improve both the system throughput time and the energy consumption by at least 10%.