Decision-Making, it's what Industry 4.0 is all about

What is Industry 4.0? Is it "smart factory'? Buying robots, automating production...is that what makes the enterprise "Industry 4.0"? Let me take a step back.

Enterprise makes all kinds of decisions during its operations: How much to sell? How much to make? What to sell? To whom to sell? Where to store? How to make?....all kind of decision and all kinds of actions followed.

To my point of view, Industry 4.0 is merely talking about how to make decisions related to productions fast and in the way applying new technologies such as IoT, Big Data (Analytics), Cloud. But in the end of day, it still comes to "Decisions".

With data collected and analyzed in real time fashion, machines can self-decide what to do next. For example, here comes a lot and machine reads its RFID to know the following operation required for this lot, (make decisions) download process recipe, start the operation (take actions). So, I think Industry 4.0 is all about how to automate the decision process without or with limited human input to improve production efficiency, reduce energy consumption, improve working environment.

Data, analysis, decisions, actions are the steps for enterprise to manage its daily operations. Because of maturity of the enterprise, enterprise might have different levels of human input during the analysis step.

If enterprise can only get "what was happening" from their data, they need high level of human input to make decisions. From my experience, production planners download production reports, manipulate those data in Excel then decide the next production plan. So, lots of human input here. This is "descriptive analytics", very basic.

Some enterprise are able to present "why did it happen" using some business intelligence tools. They can know the root cause then decide what to do next. For example, planners drill down the reports and find out one of the machines is the bottleneck, causes WIP to pile up. Planners can then decide how to offload and make sure orders still meet the due date. Less human input then stage one, this is "diagnostic analytics".

Applying some statistical models, enterprise can find out "what will happen" so they can decide to take some actions before it happens. Planners forecast incoming demand by looking into forecast numbers generated from historical data and decide the master production schedule accordingly. This is "predictive analytics".

Three stages need human inputs to make decisions, most of the time, by "experience". Therefore, it could be less efficient and hinder the responsiveness.

If some sort of "intelligence" can be applied to make decisions for human, then it can greatly improve efficiency. This is "prescriptive analytics" and is what Industry 4.0 aims for. However, there are still 2 levels of this stage, one is the "intelligence" only helps human to make decisions (decision support). For example, advanced planning system generates production plan via the data but planners still need to make final decisions for execution. The next step is really what Industry 4.0 talking about: automate decision making. Just like Jarvis in Iron Man movies, it can make the suits based on input from Tony Stark (closed color, materials) without Stark oversees the production processes.

Back to my argument, Industry 4.0 is talking about making production-related decisions in better way, faster way and automates decisions where it's possible and efficient. Therefore, to start with Industry 4.0, enterprise needs to think about how their decisions will be made then lay out the plan to enable the decision processes.