Framework to embed machine learning algorithms in P-graph: Communication from the chemical process perspectives

• Published in: Chemical engineering research & design Vol. 188; pp. 265 - 270
• Main Authors: Teng, Sin Yong, Orosz, Ákos, How, Bing Shen, Pimentel, Jean, Friedler, Ferenc, Jansen, Jeroen J.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.12.2022
• Subjects: Machine Learning; P-graph; Process Network Synthesis; Programming Interface; Python; Programming Interface; PNS; PLS; AI; P-graph; WEKA; Machine Learning; Process Network Synthesis; Python
• ISSN: 0263-8762; 1876-4800; 1744-3563; 1744-3598; 0957-5820
• DOI: 10.1016/j.cherd.2022.09.043

P-graph is a popularly used framework for process network synthesis (PNS) and network topological optimization. This short communication introduces a Python interface for P-graph to serve as a linkage to modern programming ecosystems. This allows for a wider application of the fast and efficient P-graph solver, to provide structural and topological enumeration in numerous fields. The proposed framework allows for more integrative usage in Artificial Intelligence (AI), machine learning, process system engineering, chemical engineering and chemometrics. Large and repetitive topologies can also be automated using the new programming interface, saving time and effort in modelling. This short communication serves as a demonstration of the newly developed open-sourced P-graph interface. [Display omitted] •A novel and open-sourced Python library is developed for the P-graph framework.•Allows for automatic P-graph construction and embedding of machine learning models.•Open-sourced examples demonstrate topology automation and machine learning.•The library also allows for seamless swapping between Python and P-graph Studio.•Opens path for data-driven applications to be integrated with the P-graph framework.