Ai-enabled digital twin platform for circular packaging network optimization
An AI-enabled digital twin platform that optimizes circular packaging operations using system-of-systems engineering and reinforcement learning. It models complex stakeholder interactions and logistics flows to improve network efficiency and sustainability, while avoiding local changes that could create downstream negative effects. Digital twin simulation supports adaptive optimization under unforeseen scenarios such as technology shifts and material flow variations.
Overview
The AI-enabled digital twin platform is a solution for optimizing circular packaging operations. Using system-of-systems engineering, it models complex interactions between stakeholders such as collectors, recyclers, and logistics providers. Reinforcement learning is integrated to proactively optimize operations while accounting for unforeseen scenarios, including technological changes and material flow variations. This approach aims to optimize the network globally, reducing the risk of localized improvements that could negatively impact other stakeholders or create long-term detrimental effects.
Technical specifications
- System-of-systems engineering: Models the dynamic network of stakeholders, enabling scalable and adaptable integration.
- Digital twin simulation: Simulates stakeholder behaviors and environmental interactions to support decision-making.
- Reinforcement learning: Uses digital twin simulations to develop adaptive optimization policies for unseen scenarios.
- Surrogate modeling: Enables real-time optimization capabilities without the computational burden of full reinforcement learning models.
- Collaborative information hub: Connects stakeholders to improve collaboration and provide decision support across the network.
Technology readiness level
The current development stage of this technology is at TRL 3, indicating that it has been validated in a laboratory environment. Future validation plans include defining system boundaries, developing system-of-systems digital twin models, integrating reinforcement learning for optimization, and evaluating key performance indicators at the system level using historical data and expert reviews.
About UNSW Sydney
UNSW Sydney is a comprehensive, multi‑faculty public research university spanning STEM, medicine, business and law, and the arts and social sciences. Industry connection is built into the model through a long‑running co‑op program and co‑location with a major hospital precinct in Randwick, enabling faster translation and trials. A Canberra campus deepens defence‑aligned collaboration. Research is supported by competitive awards from the Australian Research Council and the National Health and Medical Research Council, plus state programs and industry partnerships. A dedicated technology transfer office and founder programs guide IP, licensing, and startup formation with clear pathways for sponsored research.
delivered to your inbox.