Optimizing machine behavior in supply chain and industrial operations has been constrained by the inability of traditional solvers to handle complexity. Multi-objectives, multi-agents and low latency all presented problems. Deep reinforcement learning has produced significant performance gains in problems ranging from machine scheduling, order sequencing, autonomous vehicle routing, and price prediction. In this talk, Chris Nicholson (Pathmind) will discuss how these solutions learned new behavior while using Ray and RLlib.
Audience Q&A follows the presentation.