You are here
» People » Dr Alban GrastienDr Alban Grastien
I am a researcher in (Symbolic) Artificial Intelligence, primarily Model Based Diagnosis (MBD) and AI Planning. MBD is the problem of detecting and identifying defects in a physical system (such as a vehicle, a power network, or a web-service) by reasoning about the possible behaviours of the system as described by a model. In this domain, I proposed a general framework to solve diagnosis problems. I proposed a number of algorithms. I also looked at problems linked to modelling and diagnosability.
In planning, I considered the problem of conformant planning (planning despite uncertainties on actions effects and initial conditions). I also worked on path-finding, and am a co-author of Jump Point Search, Anya, and Polyanya.
I studied problems connected to power flow equations.
Through this work, I have become knowledgeable in areas such as Propositional logic (e.g., SAT, BDDs) and SMT. I am also interested in complexity and model-checking.
My main research topic is the diagnosis of discrete-event systems. The basic idea is the following. Consider a system (for instance a machine, such as a computer, a car, a space robot, or machine in a factory, etc.) which performs some actions. The system is subject to faults (such as short-circuit, leaking, break of a component, etc.) which leads to an uncorrect behaviour of the system. The goal is to use the observations (alarms generated by the system, informations provided by sensors, etc.) on the system find out what happened on the system. More precisely, I am interested in discrete-event systems. Such systems are so that their behaviour is not continuous but can be modeled as a discrete evolution (by events). A set of behaviours on such a system can be represented by an automaton. For instance, the model is often an automaton (or an equivalent representation). I often define the diagnosis as the computation of all the possible behaviours on the system consistent with the observations. This can also be represented by an automaton. The problem is quite well defined, and the main issue is the complexity which is exponential in the number components in the system.
Complete student projects
Principal investigator
Supervisor
Potential student projects
2017
Hassan Ibrahim, Philippe Dague, Alban Grastien, Lina Ye, and Laurent Simon. Diagnosability planning for controllable discrete event systems. In Conference on Artificial Intelligence (AAAI-17), pages 2481-2487, 2017.
Alban Grastien, Louise Travé-Massuyès, and Vicenç Puig. Solving diagnosability of hybrid systems via abstraction and discrete event techniques. In World Congress of the International Federation of Automatic Control (WC-17), pages 5174-5179, 2017.
Alban Grastien and Enrico Scala. Intelligent belief state sampling for conformant planning. In International Joint Conference on Artificial Intelligence (IJCAI-17), pages 4317-4323, 2017.
Michael Cui, Daniel Harabor, and Alban Grastien. Compromise-free pathfinding on a navigation mesh. In International Joint Conference on Artificial Intelligence (IJCAI-17), pages 496-502, 2017.
Cody Christopher, Yannick Pencolé, and Alban Grastien. Inference of fault signatures of discrete-event systems from event logs. In International Workshop on Principles of Diagnosis (DX-17), 2017.
2016
Daniel Harabor, Alban Grastien, Dindar Öz, and Vural Aksakalli. Optimal any-angle pathfinding in practice. Journal of Artificial Intelligence Research (JAIR), 56:89-118, 2016.
Xingyu Su, Marina Zanella, and Alban Grastien. Diagnosability of discrete-event systems with uncertain observations. In International Joint Conference on Artificial Intelligence (IJCAI-16), pages 1265-1271, 2016.
Alban Grastien, Louise Travé-Massuyès, and Vicenç Puig. Solving diagnosability of hybrid systems via abstraction and discrete-event techniques. In International Workshop on Principles of Diagnosis (DX-16), 2016.
Xingyu Su, Marina Zanella, and Alban Grastien. Diagnosability of discrete-event systems with uncertain observations. In International Workshop on Principles of Diagnosis (DX-16), 2016.
Hassan Ibrahim, Philippe Dague, Alban Grastien, Lina Ye, and Laurent Simon. Diagnosability planning for controllable discrete event systems. In International Workshop on Principles of Diagnosis (DX-16), 2016.
2015
Alban Grastien. Self-healing as a combination of consistency checks and conformant planning problems. In International Workshop on Principles of Diagnosis (DX-15), pages 105-112, 2015.
Cody Christopher and Alban Grastien. Formulating event-based critical observations in diagnostic problems. In International Workshop on Principles of Diagnosis (DX-15), pages 119-126, 2015.
Cody Christopher and Alban Grastien. Formulating event-based critical observations in diagnostic problems. In IEEE Conference on Decision and Control (CDC-15), pages 4462-4467, 2015.
Karsten Lehmann, Alban Grastien, and Pascal Van Hentenryck. AC-feasibility on tree networks is NP-hard. IEEE Transactions on Power Systems (TPS), 31(1):798-801, 2015.
2014
Alban Grastien. Diagnosis of hybrid systems with SMT: opportunities and challenges. In European Conference on Artificial Intelligence (ECAI-14), pages 405-410, 2014.
Xingyu Su and Alban Grastien. Verifying the precision of diagnostic algorithms. In European Conference on Artificial Intelligence (ECAI-14), pages 861-866, 2014.
Daniel Harabor and Alban Grastien. Improving jump point search. In International Conference on Automated Planning and Scheduling (ICAPS-14), pages 128-135, 2014.
Cody Christopher, Marie-Odile Cordier, and Alban Grastien. Critical observations in a diagnostic problem. In International Workshop on Principles of Diagnosis (DX-14), 2014.
Xingyu Su, Alban Grastien, and Yannick Pencolé. Window-based diagnostic algorithms for discrete event systems: what information to remember. In International Workshop on Principles of Diagnosis (DX-14), 2014.
Cody Christopher, Marie-Odile Cordier, and Alban Grastien. Critical observations in a diagnostic problem. In IEEE Conference on Decision and Control (CDC-14), pages 382-387, 2014.
Karsten Lehmann, Alban Grastien, and Pascal Van Hentenryck. AC-feasibility on tree networks is NP-hard. arXiv Tech Report 1410.8253.
2013
Daniel Harabor and Alban Grastien. An optimal any-angle pathfinding algorithm. In International Conference on Automated Planning and Scheduling (ICAPS-13), pages 308-311, 2013.
Franck Cassez and Alban Grastien. Predictability of event occurrences in timed systems. In International Workshop on Formal Modeling and Analysis of Timed Systems (FORMATS-13), 2013.
Alban Grastien. Diagnosis of hybrid systems by consistency testing. In International Workshop on Principles of Diagnosis (DX-13), pages 9-14, 2013.
Alban Grastien. A spectrum of diagnosis algorithms. In International Workshop on Principles of Diagnosis (DX-13), pages 130-135, 2013.
Xingyu Su and Alban Grastien. Diagnosis of discrete event systems by independent windows. In International Workshop on Principles of Diagnosis (DX-13), pages 148-153, 2013.
Alban Grastien and Anbu Anbulagan. Diagnosis of discrete event systems using satisfiability algorithms: a theoretical and empirical study. IEEE Transactions on Automatic Control (TAC), 58(12):3070-3083, 2013.
2012
Alban Grastien, Patrik Haslum, and Sylvie Thiébaux. Conflict-based diagnosis of discrete event systems: theory and practice. In International Conference on the Principles of Knowledge Representation and Reasoning (KR-12), pages 489-499, 2012.
Daniel Harabor and Alban Grastien. The JPS pathfinding system. In Annual Symposium on Combinatorial Search (SOCS-12), 2012.
Priscilla Kan John and Alban Grastien. Comparison of distributed diagnosis methods on networks with different properties. In Diagnostic Reasoning: Model Analysis and Performance (DREAMAP-12), pages 11-18, 2012.
Alban Grastien. An example illustrating the imprecision of the efficient approach for diagnosis of Petri nets via integer linear programming. arXiv Tech Report 1210.4231.
2011
Daniel Harabor and Alban Grastien. Online graph pruning for pathfinding on grid maps. In Conference on Artificial Intelligence (AAAI-11), 2011.
Patrik Haslum and Alban Grastien. Diagnosis as planning: two case studies. In Scheduling and Planning Applications Workshop (SPARK-11), pages 37-44, 2011.
Andreas Bauer, Adi Botea, Alban Grastien, Patrik Haslum, and Jussi Rintanen. Alarm processing with model-based diagnosis of discrete event systems. In AI for an Intelligent Planet (AIIP-11), 2011.
Alban Grastien and Gianluca Torta. Reformulation for the diagnosis of discrete-event systems. In Symposium on Abstraction, Reformulation and Approximation (SARA-11), pages 42-49, 2011.
Alban Grastien and Gianluca Torta. A theory of abstraction for diagnosis of discrete-event systems. In Symposium on Abstraction, Reformulation and Approximation (SARA-11), pages 50-57, 2011.
Alban Grastien. Diagnosis properties by design. In International Workshop on Principles of Diagnosis (DX-11), pages 155-158, 2011. [Open problems track]
Andreas Bauer, Adi Botea, Alban Grastien, Patrik Haslum, and Jussi Rintanen. Alarm processing with model-based diagnosis of discrete event systems. In International Workshop on Principles of Diagnosis (DX-11), pages 52-59, 2011.
Alban Grastien, Patrik Haslum, and Sylvie Thiébaux. Exhaustive diagnosis of discrete event systems through exploration of the hypothesis space. In International Workshop on Principles of Diagnosis (DX-11), pages 60-67, 2011.
