Proceedings of the

The 33rd European Safety and Reliability Conference (ESREL 2023)
3 – 8 September 2023, Southampton, UK

Dynamic Influence Diagram-Based Deep Reinforcement Learning Framework and Application for Decision Support for Operators in Control Rooms

Joseph Mietkiewicz1, Ammar N. Abbas2, Chidera Winifred Amazu3, Anders L. Madsen4 and Gabriele Baldissone5

1Hugin Expert A/S /TU Dublin, Denmark. Food Science Environmental Health, Technological University Dublin, Ireland.

2Department of Computer Science, Technological University Dublin, Ireland. Data Science, Software Competence Center Hagenberg, Austria.

3Safety, Reliability and Risk Centre (SAfeR), Department of Applied Sciences and Technology, Politecnico di Torino, Italy.

4Hugin Expert A/S, Department of computer science Aalborg University, Denmark.

5SAfeR - Department of Applied Sciences and Technology, Politecnico di Torino, Italy.


In today's complex industrial environment, operators are often faced with challenging situations that require quick and accurate decision-making. The human-machine interface (HMI) can display too much information, leading to information overload and potentially compromising the operator's ability to respond effectively. To address this challenge, decision support models are needed to assist operators in identifying and responding to potential safety incidents. In this paper, we present an experiment to evaluate the effectiveness of a recommendation system in addressing the challenge of information overload. The case study focuses on a formaldehyde production simulator and examines the performance of an improved Human-Machine Interface (HMI) with the use of an AI-based recommendation system utilizing a dynamic influence diagram in conjunction with reinforcement learning. The preliminary results indicate the potential of these methods to aid operators in decision-making during challenging situations and enhance process safety in the industry.

Keywords: Decision support, Dynamic influence diagram, Reinforcement learning, Process safety, Intervention procedures, Human-machine interface.

Download PDF