![anylogic hack anylogic hack](https://next-hack.com/wp-content/uploads/2020/02/Levels.png)
RL is no more impossible than other ML approaches, based on that argument.Ī valid simulation of a physical system is a source of synthetic data that helps solve that problem.ĭeep RL is already controlling machines on factory floors, and it will slowly help optimize larger and larger systems, from groups of assembly lines, to networks of plants.
![anylogic hack anylogic hack](https://electrek.co/wp-content/uploads/sites/3/2016/02/tesla-xray-2-e1455320423724.jpg)
It's messy, partial, and usually not gathered with ML in mind. The problems with data are pretty much the same problems you get with ML in general. I see a lot of preconceptions about RL in this thread that are partially true, but also solvable. This is the source of many of the gains that RL produces. Solving the third problem gives rise to very interesting emergent behavior among teams of machines, which learn to behave in ways that are almost impossible for a hard-coded set of rules to specify. * multiple agents making simultaneous decisions in coordination * multiple objectives in complex scenarios Yes, these show RL applied to simulations, and yes, real data and real physical plants are more complex than that.īut those physical systems are already controlled by optimizers (usually mathematical solvers like IBM Cplex or Gurobi), and deep RL happens to be an optimizer that can handle a few things better:
![anylogic hack anylogic hack](https://images.squarespace-cdn.com/content/v1/5521673ce4b09eb81bf68349/1440959030310-DFZWWDZQ4IOTE4BVW4VC/image-asset.png)
Our team at Pathmind has applied deep RL to multiple use cases in industrial control and supply chain management, notably various forms of scheduling, and MEIO, respectively.