SwarmIn – Optimizing Energy and Resource Efficiency

SwarmIn’s innovative approach integrates AI to optimize production plants by balancing work-in-progress (WIP) and flow factors with energy and resource efficiency. Our strategy combines high-level combinatorial optimization with low-level swarm intelligence, ensuring a novel synergy between cyber-physical systems and human agents to enhance industry 4.0 sustainability. This dual-layer optimization frames the plant as a self-organizing system, fostering unprecedented operational efficiency.

Partner

Lakeside Labs (Coordinator)
Melanie Schranz, Lakeside Labs B04b, 9020 Klagenfurt
https://www.lakeside-labs.com/

Infineon Technologies Austria AG
Walter Laure, Siemensstraße 2, 9500 Villach
http://www.infineon.com/cms/austria

University of Klagenfurt
Martin Gebser, Universitätsstraße 65-67, 9020 Klagenfurt
https://www.aau.at/en/aics/

University of Graz
Nina Hampl, Merangasse 18/EG, 8010 Graz
https://umweltsystemwissenschaften.uni-graz.at/de/

Messfeld GmbH
Michèle Posch
Lakeside B07a, 9020 Klagenfurt
https://www.messfeld.com/

Novunex GmbH
RTD, Völkermarkter Ring 7-1, 9020 Klagenfurt

Optimizing Energy and Resource Efficiency in Industry 4.0 Through Advanced Cyber-Physical Systems

The pursuit of climate objectives, alongside the challenges of escalating energy demands and previous struggles with supply chain stability, is driving the industrial sector to prioritize finding efficiencies in energy and resource usage for the foreseeable future. The framework of sustainable Industry 4.0 and cyber-physical systems in manufacturing plays a pivotal role because it not only facilitates real-time interactions, utilizes machine learning (AI), and simplifies monitoring activities but is also designed to boost energy and resource efficiency through its capacity to offer instant data via edge computing. Additionally, the increasing variety of products and the historical expansion of industrial plants add layers of complexity, making the scheduling tasks exceedingly difficult, classifying them as NP-hard. Taking a semiconductor factory as a case in point, the scheduling involves coordinating activities across 400 to 1,200 different stations, where over 1,500 distinct products go through approximately 300 varied processing steps. Here, linear optimization techniques falter, unable to handle the vast, intricate, and dynamic nature of the required computational tasks within a reasonable timeframe. These strategies are thus limited to smaller sections of the plant and fail to consider holistic system behaviors, leaving significant optimization potentials untapped.

Project Details

The project, titled “SwarmIn: Swarm Intelligence and Combinatorial Optimization for Energy Efficient and Adaptive Industry 4.0,” kicked off in 2022 and is planned to conclude on August 31, 2025. Spanning a total of 36 months, SwarmIn focuses on enhancing energy and resource efficiency within the context of Industry 4.0 through advanced AI-based methodologies. The project aims to integrate swarm intelligence, a dimension of AI, with combinatorial optimization techniques to address the complexities of modern industrial environments such as semiconductor factories that feature extensive product diversity and intricate process dynamics. This innovative approach looks at a synergetic use of central and local optimizations to enable real-time, efficient, and adaptable production systems.

Project Goals

By the project’s end, it is expected to yield a suite of algorithms and software libraries that significantly advance the capabilities of cyber-physical systems and human-machine interfaces in industrial settings. The introduction of humans as swarm agents in this optimized ecosystem is a novel concept, aiming to contribute significantly to the domain of AI and industrial system optimization through scholarly publications and dissemination at prominent conferences and journals. This body of work will affirm the project’s pioneering status and its impact on advancing sustainable industrial practices.

Introducing SwarmIn: Pioneering AI-driven Optimization for Sustainable Industry 4.0 Production

In the SwarmIn initiative, the primary objective is to achieve a harmony between work-in-progress (WIP) levels and flow efficiency, while also optimizing the production facility to enhance energy and resource utilization. Our strategy involves crafting a pioneering architecture that marries various artificial intelligence (AI) techniques to leverage their strengths effectively: initially, we utilize combinatorial optimization to globally estimate configuration parameters, effectively narrowing down the solution space. This refined space then serves as the foundation for the subsequent, more granular optimization, which employs swarm intelligence. This method adopts a collective, grassroots approach through multi-agent systems. This innovative blend of approaches is unique, especially in the context of semiconductor manufacturing, as it integrates both cyber-physical elements (like machinery and batches) and human participants as agents influencing energy and resource efficiency within an Industry 4.0 framework. At the ground level, each agent operates according to a specific set of localized rules. The synergistic link between the overarching and detailed optimization levels introduces a novel concept, transforming the production environment into an autonomous system of collaborating agents. The detail-oriented optimization leans on swarm intelligence, a niche within AI, and is executed close to the operational core (machines and batches), while the broader combinatorial optimization is managed centrally, potentially through cloud services that integrate seamlessly with the factory’s ERP system. These two optimization layers maintain a dynamic interplay via (wireless) communication networks, with 5G technology potentially serving pivotal roles in meeting Industry 4.0’s demands for low latency and high reliability. Another key innovation involves the integration of energy and resource efficiency metrics, measured through retrofit sensors in the manufacturing process and human feedback, into both optimization stages. This not only aids in analyzing these metrics but also embeds them into the decision-making processes for CPS (cyber physical system) and human-machine interactions, fostering an adaptive, efficiency-oriented behavior. Ultimately, SwarmIn aims to deliver a suite of algorithms and simulations for both macro and micro-level optimizations to address the intricate complexities of Industry 4.0 manufacturing and meet stringent energy and resource efficiency standards. This suite will be accessible as a software library. Setting itself apart from existing efforts, SwarmIn devises a ground-breaking architecture that combines disparate AI strategies (combinatorial optimization and swarm intelligence) for the first time, based on their distinct attributes. Further enriching this architecture is the inclusion of human actors as integral members of the swarm and incorporating energy and resource efficiency as key dimensions in shaping the sustainable production facilities of the future.

Production of the Future

Production of the Future is the funding program of the Federal Ministry for Climate Action, Environment, Energy, Mobility, Innovation and Technology (BMK). Funding is provided for cooperative projects and flagship projects of industrial research or experimental development, exploratory projects (pilot study for R&D project), endowed professorships, individual projects of experimental development and knowledge transfer under the FFG’s BRIDGE program.

A Comprehensive Perspective

The call aims to contribute to the goal of the federal government to lead Austria from the group of so-called Innovation Followers to the group of Innovation Leaders – to become one of the most innovative countries within the EU. The initiative is aimed at companies, research and technology organisations, universities, universities of applied sciences and secondary technical colleges based in Austria.