Datadrevet menneske-i-loopen optimeringsarbeidsflyt

Offentlig prosjektsammendrag

In this project we will build the next version of our ProductionCompas AI (PCAI) software, extending it from its current state — a set of problem-specific applications aimed at the oil and gas industry — to an AI assistant which provides intelligent suggestions in real-time for industrial assets in a broad range of industries. To achieve this, we will re-engineer our algorithms and applications to provide action-oriented solutions to relevant subproblems, and create an integrating layer which combine these to solve the multi-objective optimization problem of asset-level decision making. The application of this framework to the fish farming industry will serve as an important benchmark problem for the overarching objective of this project: A human-machine optimization workflow which integrates industry-tailored applications and provides action suggestions which adapt according to human feedback over time, leading to decisions being made in agreement between human operators and PCAI.