Using Simulation to Train and Test Artificial Intelligence for Business Applications


Originally presented at the GE EDGE & Controls Symposium 2019 by Dr. Andrei Borshchev. This presentation details why simulation and artificial intelligence work well together. It explores how simulation benefits AI through the generation of synthetic data, the provision of learning environments, and the creation of testbeds for trained AI. Dr. Andrei Borshchev introduces the methods for integrating AI and simulation before demonstrating a technical example and an industrial case study.

  • Introduction
  • What is simulation modeling? What is AnyLogic?
  • Digital Twins and Case study: Gas Turbine Fleet
  • Why AI and Simulation? Some AI terminology. How can simulation help AI?
  • Example of Deep Reinforcement Learning using simulation: Traffic Light Control (detailed)
  • Case study: AI trained by simulation model in Ferromagnetic Core Production
  • Conclusion. What are the challenges?

The presenter, Dr. Andrei Borshchev, is the founder and CEO of The AnyLogic Company and the author of the Big Book of Simulation Modeling. He leads the AnyLogic software roadmap and has introduced features that have become standard in the simulation society, such as multimethod modeling, agent-based simulation linked with GIS maps, and cloud model execution.

관련 영상