Simultaneous Production and AGV Scheduling using Multi-Agent Deep Reinforcement Learning
Jens Popper, Vassilios Yfantis, Martin Ruskowski
Abstract
Increasing demand for customized products in the wake of the 4th Industrial Revolution is placing ever increasing demands on the flexibility of manufacturing systems. Furthermore, the increasing usage of automated guided vehicles (AGV) adds another layer of flexibility and also complexity to the overall production system. The resulting Flexible Job Shop Scheduling Problem (FJSSP), including the coordination of the AGVs, is NP-hard and therefore hard to optimize. To address this problem, a Reinforcement Learning Multi Agent (MARL) system is proposed, in which job scheduling and vehicle planning is done cooperatively. This concept is described and prototypically implemented.
Topics & Concepts
Reinforcement learningScheduling (production processes)Job shopFlexibility (engineering)Job shop schedulingComputer scienceIndustrial engineeringFlexible manufacturing systemEngineeringManufacturing engineeringFlow shop schedulingDistributed computingArtificial intelligenceEmbedded systemOperations managementRouting (electronic design automation)MathematicsStatisticsScheduling and Optimization AlgorithmsAdvanced Manufacturing and Logistics OptimizationFlexible and Reconfigurable Manufacturing Systems