How Holonic Dreams are Becoming Manufacturing Realities

Let me tell you, there are few things more thrilling than seeing a concept you poured your heart into decades ago slowly coming to life in ways you barely dared to dream. For me, that feeling hits hard with the incredible capabilities of AI in general, and specifically, Generative AI . These aren't just buzzwords; they're fundamentally reshaping how we think about digital manufacturing (or whatever the latest term is). While Generative AI is phenomenal for creating content, designing new products, or even simulating complex processes, its true power in manufacturing often lies in its ability to empower something even more profound - Multi-Agent systems. This is the realization of Holonic concepts in a composable manner to enable agile manufacturing with Agentic AI .

When I see these advancements I am just mesmerized and my mind goes back to the 1990s. That’s when my journey into this future really began, deeply immersed in the world of Holonic Manufacturing Systems (HMS) and the emerging field of multi-agent systems.

Back then, I was a young researcher PhD working a methodology and architecture for Holonic manufacturing systems in collaboration with other like minded researchers as part of global consortium. It wasn’t just an academic exercise; it was a burning ambition to make manufacturing truly agile and resilient. I envisioned a factory floor that wasn't a rigid, top-down hierarchy, but a vibrant, decentralized network of intelligent, collaborative entities – what we called "holons."

The Holonic Vision: A Glimpse into the Future I Believed In

Imagine a shop floor where every machine, every production cell, every product, wasn't just a passive component but an intelligent "holon" – a self-contained, self-regulating unit. They would have their own smarts, making decisions, talking to each other, and collectively adapting to whatever curveball the market threw at them. Koestler's concept of a "holon" – simultaneously a whole and a part – perfectly captured this idea of distributed intelligence.

The benefits? Oh, they were clear and seemed a world away but yet an eerily anticipatory need of the current political and economic circumstances.

• Agility beyond belief : Reconfiguring production lines in a flash, launching new products on a dime, responding to customer demands with unprecedented speed.
• Built-in resilience : If one holon stumbled, the others would dynamically pick up the slack, re-routing operations to keep things flowing. Downtime issue would be a distant memory.
• Seamless scalability : Adding new machines or processes would be like plugging in a new module, effortlessly integrating into the intelligent network.
• Optimization from within : Local decisions by these smart holons would ripple up to optimize the entire system, far surpassing anything a central, rigid control system could ever hope to achieve.

This wasn't just theory; it was a blueprint for a manufacturing revolution.

We are Still Waiting for a Digital Manufacturing's Breakthrough

The holonic concept was the perfect architectural dream, but the engine to power it, multi-agent systems, was still in its infancy. My research in the 90's focused on how to design these agent systems, how to give them that holonic spirit, but the reality was that our ambition ran ahead of the available technology. We were hitting walls. The computational power needed to run complex agent logic on shop-floor controllers was simply astronomical for the time. Getting a multitude of agents to communicate reliably and securely across a factory? That was a networking nightmare. And then the AI capabilities were nascent, a whisper of potential, compared to the what we wield in our hands today."

Those were exhilarating times for pure research, pushing the theoretical limits of what manufacturing could be. But bringing it to large-scale industrial reality? That was a bridge too far. Until now ...

Today's Reality: AI and Digital Infrastructure can Unleashing the Holonic Dream

Fast forward to today, and the technological landscape has dramatically evolved. The convergence of several critical advancements has not only rendered the holonic vision achievable but has propelled it into operational realms previously unimaginable:

1. Advanced AI and Machine Learning Capabilities : Modern Artificial Intelligence and Machine Learning algorithms now provide the sophisticated analytical and cognitive capabilities for individual software agents. These algorithms enable agents to learn from large-scale industrial datasets, execute robust predictive analytics with high precision, and achieve adaptive process optimization in real-time. This represents a fundamental shift from deterministic, rule-based systems to dynamic, self-optimizing intelligence.
2. The Industrial Internet of Things (IIoT) as the Network of Agents, Powered by Agentic AI : This is where the core holonic vision finds its full realization. The Industrial Internet of Things (IIoT) is not merely a collection of connected devices; it forms the very network of agents. The IIoT nodes themselves are designed as autonomous, intelligent agents . Each smart device, each sensor, each piece of equipment can be made to act as a data acquisition point and, crucially, as a decentralized intelligent entity. The emergence of Agentic AI imbues these IIoT nodes with advanced capabilities for complex task decomposition, strategic planning, execution, and critical self-reflection. This synergy of IIoT as the foundational agent network and Agentic AI providing the cognitive layer directly mirrors the autonomous and cooperative principles fundamental to our original holonic concepts, enabling seamless interoperable collaboration across the manufacturing ecosystem. 

Futuristic Use Cases Becoming Reality: Empowering the Digital Twin for Tomorrow's Factory

For all these visions to truly become reality, we needed more than just powerful tech; we needed a concept that can effectively by applied to the realities of a physical operation. Here we can use a Digital Twin approach to represent the reality of the operation and build a foundational model with a set of constructs that aligned perfectly with the holonic principles I’d spent years researching.

This is also where Agentic AI enters the stage, becoming the intelligence that breathes life into this digital replica. My research from the 90s proposed a consistent model for these foundational constructs, embodying them as intelligent Product, Order, and Resource agents (or holons). This was the blueprint for how the real-world manufacturing elements could become autonomous, cooperative entities within the digital twin.

Let's step into this future for a moment, and I’ll paint a picture of what a manufacturing operation truly looks like when its digital twin is powered by these agentic principles, or in other words a next generation paradigm shifting composable manufacturing system . In this world, a work order isn't just data on a screen; it instantly awakens an intelligent Order Agent within the digital twin - or plainly the manufacturing system. This agent immediately gets to work, dynamically negotiating with Resource Agents —the digital representations of machines, tooling, even the specific human expertise required—to secure optimal production slots and materials in the virtual space, which then collaboratively drives actions in the physical factory assisting operators in orchestrating the operation.

As raw materials enter the facility, each component, or even the nascent product itself, manifests as a Product Agent within this digital twin. This agent carries its own unique digital identity and manufacturing instructions, literally guiding operators to route its counterpart through the physical production line. It's constantly communicating its status and needs within the digital twin, ensuring it receives the precise processing at each stage in the real world. If a specific machine (a Resource Agent ) suddenly reports a slight anomaly – say, a bearing starting to warm up – its built-in intelligence within the digital twin instantly flags it. Instead of waiting for a catastrophic failure, the line's collective intelligence, orchestrated by operators using the various software agents operating within this digital twin environment, might subtly re-route the physical Product Agent to an alternative, readily available machine. Or, better yet, the affected Resource Agent might even initiate a precise, self-healing routine or schedule a just-in-time, predictive maintenance intervention, ensuring that the issue is resolved before it impacts production, all while the Order Agent helps ensure deadlines are still met.

Quality control isn't a post-production check; it's baked into every micro-decision. Product Agents and Resource Agents , leveraging their digital twin data, are continuously monitoring parameters, spotting the tiniest deviation, and triggering immediate corrective actions to ensure "right first time". This seamless, autonomous orchestration – where products find their way, orders fulfill themselves, and machines manage their own well-being, all empowered through the precise, real-time fidelity of their digital twins – transforms the factory into a living, breathing, self-optimizing organism. It’s a level of agility, efficiency, and resilience that felt like pure science fiction in the 90s, but is now can become our tangible reality.

This leads us to an operational reality where the use cases that may seem futuristic are in fact possible , for example:

• Self-Optimizing Production Lines : Imagine entire lines monitoring themselves, sniffing out bottlenecks, predicting breakdowns, and then autonomously re-routing production or tweaking parameters to keep output optimal. Empowering human operators with currently unimaginable support in orchestrating operations.
• Dynamic Resource Allocation : Agents negotiating for machines, tools, and materials in real-time, ensuring every asset is utilized perfectly, eliminating idle time. Elevating scheduling and dispatching to unheard of levels of effectiveness and accuracy. 
• Predictive Maintenance and Self-Healing Systems : No more waiting for a breakdown. Agents predict failures with incredible accuracy and can even kick off self-repair routines or proactive maintenance, slashing downtime and costs.
• Enhanced Quality Control : Agents tirelessly monitoring processes and product quality, spotting deviations instantly and triggering immediate corrective actions. This is "right first time" manufacturing, every time.
• Boosted Compliance : Automated data collection, precise procedure execution, immutable digital records – agentic systems dramatically reduce human error and guarantee adherence to the toughest regulations.
• Unparalleled Traceability : Every single action, every decision by an agent, meticulously recorded. Audit trails become pristine, investigations swift and clear.
• Driving "Right First Time" : By minimizing variability and providing real-time feedback, these systems help ensure products meet quality specs from the outset, slashing costly rework.
• Accelerated Innovation : With more efficient, reliable processes, companies can pour more resources into R&D, bringing life-saving drugs or mission-critical components to market faster.

The journey from the elegant theories of holonic manufacturing systems to the practical, jaw-dropping capabilities of Agentic AI has been long, but intensely rewarding. What started as pure academic curiosity, exploring the power of decentralized, intelligent control, has now become the very bedrock of digital transformation in manufacturing. We’re no longer just imagining; we are actively building factories that learn, adapt, and optimize themselves , powered by the incredible, collaborative intelligence of software agents, physical machines, and empowered humans. It's a testament to the enduring power of fundamental research, a strategic commitment to true digital transformation, and the relentless, accelerating pace of technological innovation.

But hold on, not so fast. Here’s where my passion often turns to frustration. We have the technology today, and more is coming fast – innovation is accelerating at an exponential rate! Yet, a fundamental problem persists in manufacturing: so many companies still don't grasp that adopting digital technology demands a profound transformation , not just a simple upgrade.

The Agility Forum , a 1990 initiative to transform manufacturing, proclaimed that we need to thrive in an environment where change is the only constant. In today's volatile business landscape – marked by unprecedented geopolitical shifts, rapid market fluctuations, and increasingly fragile global supply chains – this is not longer a theory; it's the raw truth of survival. While our research in the 90s certainly anticipated a future of greater dynamism and the critical need for manufacturing systems to adapt , even we couldn't have fully foreseen the sheer velocity and breadth of the disruptions we face now.

This intense, continuous flux means leveraging that change, embracing agility itself, as your ultimate competitive advantage. The sheer ability to adapt, to pivot swiftly, and to continuously evolve your operations is precisely what will differentiate leaders from those left behind. In this dynamic landscape, digital transformation is not a static one time event, nor is it a project with a defined end; it is a continuous process of adaptation to changing business and technological environments - hence we need to start talking about Continuous Transformation .

The core issue isn’t the lack of innovative tools; it’s the mental, organizational, and cultural shift required to truly embrace them - yes its still really about people . Real digital manufacturing means rethinking everything – your processes, your workflows, even how you do business. It’s a complete reimagining, and that's precisely why the original holonic concepts, now enabled by modern tech, offer the inspiration for a breakthrough path. Embrace this paradigm shift, or risk being outmaneuvered by those who do. The future of manufacturing is intelligent, interconnected, and increasingly autonomous, built directly on the visionary concepts laid down decades ago.

To Data Model or not to Data Model

The ongoing debate about where and how MES fits in new era of digital technologies is raging. Its not surprising and in fact to be expected in any kind of change, basically the old guard vs the new guard. Of course you have to believe that the 4th industrial revolution is really a paradigm change. Something that I clearly align with and have some background to do so since I have been studying this phenomena since the 1990s.

As in other paradigm shifts there will always be a bit of the old that is part of the new. Steam power has not completely disappeared, it still relevant in specialized application but it is not the main source of energy powering industrial operations. This leads us to ISA-95 that I believe is a relic of the current "industry 3.0" era and not directly relevant in the new digital paradigm. (note I purposefully am trying to minimize the use of "Industry 4.0" since it starting to get a negative connotation with all the hype going on). But, that being said there are elements of ISA-95 and other best practices that may be relevant in the new paradigm, ie the old in the new?

If we let history be our teacher we can probably come up with some prediction and that is where the data model topic is interesting. ISA-95 includes a data model and all the established MOM solution include a data model that based on the available technologies at the time that seemed appropriate. The question is then; is the quest to achieve the nirvana of one standard monolithic data model for all manufacturing achievable and is it still relevant with the new digital technologies? The answer I think is clearly no and no, as far as I know there are very few, if any, examples of an organization achieving a real working standard data repository for all its operation and its not because of lack of trying.

The bottom line here is that striving for a single standard data model in a monolithic repository is a fools errand, regardless of if we try to implement it with modern digital technologies. That being said a common, shared and interpretable view of manufacturing operations is still needed and critical. In fact it's at the core of Industry 4.0, in that its the data and information that gives us the Visibility, Transparency, Predictive Capacity and Adaptability. This holistic view into the manufacturing operations is also at the core of the CIM concept from the 80s that advocated a common "shared knowledge" that all operational activities in plant uses in order to streamline to manufacturing of products. That means that both paradigms are aligned around the same challenge that to improve manufacturing operations we need to all have a common understanding and view into the operation!   

The CIM Enterprise Wheel (c)1993, SME.

The difference is how we achieve this common and shared view (information and knowledge). In the old paradigm it was the notion of a strict and rigidly structured data model, in the new paradigm we have relaxed these requirements to allow for analysis from both structured and unstructured data. I can hear the skeptics already; how can you gain any insights with different solution each having their own data structures? A few things to consider here: We do need context and this context should be defined at the source. We need to simplify data structures and get away from multiple levels of abstractions needed to run monolithic process driven solution.  Adhere to some simple shared guidelines using a consistent data dictionary that allows for flexibility within your organization. (I know this sounds overly simplistic and see part II of this blog post). With these principles we can adopt many of the modern digital tools to curate views into our data, on demand with the flexibility needed for common and personalized views and insights including of course AI.     

Let's take at an example where different solutions all represent some data about a lot of materials and its product code. The material can be referenced as Lot, Batch, Units, Pack, Kit, etc and the product code can be references as SKU, Item ID, Product, Material Number, etc. We of course immediately recognize these different names as similar because we understand how they are used. In the old paradigm we had to enforce strict rules in structure and semantics for software solution in order to visualize and analyze this data. That is however changing with new digital technologies and modern analytics platforms.  

It is also where AI can help, you see simply put AI is good at finding patterns. Its not that AI understands what the meaning of Item and Material Number is. It simply is looking for similarities in the relationship to other data structure and how its used to see that Item and Material Number really are very similar. With enough data volume and variety this can be easily detectable. Notice I said volume and variety this is where Cloud based system are important. Using isolated traditional monolithic system data sources will never get you to this point, even if they are lift and shifted to the cloud. You need a modern cloud native operational platforms that provides easy access to the their data that can be amassed and used to identifying these patterns.

I know there are a number of concepts discussed here and there may be some lack of depth in the discussion. I promised a follow up on this post with some more detail. But assuming this is true, just think about it. It means we can relax the strict data type and structure requirements and allow citizen developers to extend template data structures to create solution to solve operational problems and know that we can still gain valuable insights about operations, and again the more data we have to more insight we have. The conclusion here is: prioritize data volume and variety and not monolithic structures.

My head is in the Cloud

I have spent the last 2 months coming up to speed with SaaS, The Cloud, Big Data and advanced analytics. Wow, I am amazed at how far technology has come. In the last 7 years it seems I have moved further and further away from software technology with my increasing operational responsibilities. I did my best to follow the general trends and progress but really never in enough depth. I knew there was something “big” happening but kept thinking “I know what is going on…”, “I should have a general understanding…”. When I stopped to really understand how far some of these technologies, we group under the “Industry 4.0” umbrella, have come I found it quite astounding!

The power of artificial intelligence (AI) and machine learning (ML) algorithms that are able find patterns we never knew existed in vast amounts and use this information to predict future behavior is amazing. Stories such as the one where Target (the retail store) was able to identify a girl as pregnant before her dad knew, stand as proof. I have also found the ease at which data can send data to the Cloud is just hard to understand. We have been fighting the connectivity and data contextualization problem for such a long time that its hard to believe it can be done any other way. If we take a breather, step out of the ditch and take a look around then we can quickly realize - we don’t need to do this anymore! Data can be collected by software agents running on everything from a server to a sensor and configured through a digital twin. There is no need to format, convert, transform or contextualize the collected data. In fact the AI and ML algorithms work best if we keep it unstructured and the smaller the data sets the better.

This is all very exciting and the more I think about how we can use these technologies the more more encouraged I am that we can achieve true manufacturing intelligence. The ability to have a real time picture of everything that is going on in our manufacturing plant is simple and achievable. On top of that we can discover relationships between processes and artifacts that we never imagined existed, and use these to improve, to better operate, and maybe stop the great firefight that is operating a plant. We can transform the daily operational meetings to something akin of a weather forecast - wouldn’t that be something to watch?