A Proposal and Evaluation of a Digital Twin Framework for PI-Hubs using Re-enforcement Learning based Multi-Agent Systems Model
Author(s)
Vijay, Anshul
Thompson, Russell G.
Nassir, Neema
Advisor(s)
Editor(s)
Collections
Supplementary to:
Abstract
The Physical Internet (PI) provides a way to enhance logistic network performance spanning the social, financial, and environmental domains. Hyperconnected City Logistics (HCL) encompasses logistic activities arising within PI across the greater metropolitan region. To enable PI to operate as an open, collaborative network, a standard form of data exchange amongst network participants is required.
GS1 Scan 4 Transport’s Digital Link (GS1DL) is a data sharing standard enabling tracking of products through the supply chain. This includes the realization of shorter and more consistent processing times at goods transfer points within facilities. The potential impact of utilizing the GS1DL within a collaborative environment is yet to be investigated. Digital Twin (DT) enables real-time monitoring of assets’ statuses and tracking of containers, leading to a real-time virtual representation of the physical facility, integrating the various computer network systems.
This study proposes a novel Digital Twin framework for PI-Hubs, integrating a re-enforcement learning based multi-agent system (MAS). Real-time location of goods flowing through the facility will be used by machine learning to predict the likelihood of containers arriving at the outbound docks, to be subsequently re-allocated to outbound vehicles via a reallocation optimization algorithm. The evaluation of the DT framework on PI-Hub's operational performance will include the assessment of Scan4Transport’s Digital Link standard.
Sponsor
Date
2023-06
Extent
Resource Type
Text
Resource Subtype
Paper