Artificial Intelligence / AI – Smart Factory Glossary

In general terms, artificial intelligence, or AI for short, is the ability of a machine to act and think like humans, and to perform tasks that require human intelligence. According to a definition by the European Parliament, AI imitates human abilities such as reasoning, learning, planning, and creativity to do so. AI uses and combines different methods and technologies to approach its goal. These methods and technologies include machine learning, cognitive modeling, pattern recognition as well as image and language processing. AI systems have taken root in a wide variety of fields and industries to solve complex tasks independently and without human intervention.

 

Types of artificial intelligence

From a scientific point of view, the term artificial intelligence is not clearly defined. However, leading researchers do agree on a rough differentiation between weak and strong AI.

• Weak AI means that AI supports people with specific tasks and issues to achieve the best possible result.

• Whereas strong AI independently performs any task. To do this, AI trains itself in new skills, learns new technologies, and creatively applies them. Thus, strong AI reaches or even surpasses the level of human intelligence. In some future scenarios conceived by human imagination, artificial intelligence even develops its own consciousness. Today's AI is far from that: all current AI applications are examples for weak AI.

   

AI in the production environment

In the production environment, AI is used to evaluate large amounts of data and reveal correlations. The more historic data and influencing factors are analyzed, the more precise are the predictions made and production efficiency can be increased accordingly. AI predictions that are based on the analyzed data are incorporated in production planning: Predictive Maintenance is a common example. Predictive Maintenance integrates all machine and production data to purposefully plan machine maintenance. This avoids unplanned breakdowns that could lead to delays in delivery. Detailed scheduling also benefits from artificial intelligence. The manufacturing app AI Planning by MPDV that is part of the Advanced Planning and Scheduling System FEDRA uses reinforcement learning to plan machines and operations in the best possible way.

MPDV's AI-based scrap analysis is another example. The application identifies the factors that lead to a good or bad scrap rate. AI integrates order and article data to evaluate different combinations of material, machine, tool, and color. Finally, AI identifies which combinations result in high or low scrap rates. In the next step, the user's domain knowledge is required: it's up to the user to decide which parameters should be changed to reduce the scrap rate of a specific combination.

In brief, artificial intelligence supports production with the following requirements:

• Planning orders
• Analyzing real-time data
• Predicting events