Swarm Intelligence and Combinatorial Op...

In SwarmIn the overall goal is to balance the WIP waves and flow factors by applying a production plant optimization featured with energy- and resource-efficiency parameters in an industry 4.0 setting as given with a production plant of, e.g., IFAT. These parameters will enrich the optimization procedure that builds upon a tightly interconnected high-level, combinatorial and a low-level swarm intelligence optimization that presents the project’s output in form of an artificial intelligence (AI)-based software library. In combinatorial optimization the focus lies on global production task planning, resource allocation, and cyber-physical system (CPS) configuration, yet based on abstract models of the production process chain to manage the complexity and uncertainty of industrial CPSs. In optimization using swarm intelligence, the algorithms are considered as local rules for the agents (heterogeneous swarm of CPSs and humans) performing the optimization from bottom-up. We design a novel mixed AI architecture, where the low-level optimization with swarm intelligence runs on local edges supported by retrofit sensors and local data reduction as input for the local decision rules, whereas high-level optimization is done centrally, e.g., as a cloud service.