@ARTICLE{10.21494/ISTE.OP.2019.0403, 

TITLE={Diagnosis of partially observed Petri nets with algebraic indicators}, 
  
AUTHOR={Amira Chouchane, Philippe Declerck, }, 
	
JOURNAL={Industrial and Systems Engineering}, 
	
VOLUME={2}, 

NUMBER={Issue 1</br>}, 
	
YEAR={2019}, 

URL={https://openscience.fr/Diagnosis-of-partially-observed-Petri-nets-with-algebraic-indicators}, 
	
DOI={10.21494/ISTE.OP.2019.0403}, 
	
ISSN={2632-5993}, 
	
ABSTRACT={In this paper, we present a diagnostic approach for partially observable Petri nets based on the formalization of fault indicators under the algebraic form (min, max, +). The faults are modeled by certain unobservable transitions, and the occurrence of a fault is equivalent to the firing of the associated transition. The Fourier-Motzkin elimination algorithm is applied off-line for the generation of fault indicators used on-line for fault diagnosis in a polynomial time. This approach allows anticipating the system state (faulty, no-faulty, uncertain) for each observed transition of the observation. A comparison with the diagnostic technique using the classical form of linear programming problems shows the effectiveness of our approach. The diagnostic system is a module operating in parallel with the classical approach.}}