Smart tech for real-time industrial planning
These are uncertain times for industry, as it navigates survival with geopolitical changes, inflation, and supply chain issues.
Chemical engineering researchers are developing innovative methods to harness machine learning (ML) for industrial applications, helping industries plan production more effectively in the face of unpredictable conditions.
A research group led by Professor Luis Ricardez-Sandoval is using ML methods to train “smart agents” to make production scheduling decisions in chemical and manufacturing systems where there is uncertainty.
The agents are trained through simulations of plant processes that include unexpected events, for example, equipment failure or a sudden change in production demands.
“We train the agents to provide scheduling decisions in real time that adjust to the uncertainty of the process, avoiding the need to solve complex optimization problems every time there is an unexpected event so we can ensure that we’re deploying an attractive production schedule in real-time” says Ricardez-Sandoval a Canada Research Chair in Multiscale Modelling and Process Systems Engineering.
The research group uses Reinforcement Learning to train the agents to make scheduling decisions in real time.
The novelty of the method is that the agents are designed with special memory layers called long short-term memory (LSTM), which helps them remember events that took place in the past (historical data). Agents are designed to make multiple decisions simultaneously. When tested against an agent that did not have LSTM, the proposed method performed more efficiently.
“The agent’s brain is modelled using Artificial Neural Networks, and we used the design of networks from other areas like robotics and natural language processing to create a special agent that can handle the uncertainty in industrial processes,” said PhD student Daniel Rangel-Martinez, who had a leading role in this project.
The research group uses Reinforcement Learning to train the agents to make scheduling decisions in real time. These techniques promote learning by receiving feedback from actions. To ensure the agents learn effectively, researchers use a training approach that instructs them to focus on the main task even when things change or go wrong.
Testing on two industrial plants concluded that combining memory with Reinforcement Learning is effective in managing complex large-scale scheduling problems.
A paper on their work titled Recurrent Reinforcement Learning Strategy with a Parameterized Agent for Online Scheduling of a State Task Network Under Uncertainty | Industrial & Engineering Chemistry Research was published in the journal of Industrial & Engineering Chemistry Research.