At ANU, I am a member of the Planning and Optimisation research group in the School of Computing. In Toulouse I am a member of the Artificial and Natural Intelligence (ANITI) institute and the RIS group at the LAAS-CNRS Laboratory.
My research interests are in automated planning, scheduling, diagnosis, and search, their integration with optimisation, machine learning, and verification, as well as their applications to energy and transport.
I am a AAAI Fellow and Co-Editor in Chief of the Artificial Intelligence journal.
My educational background includes a Diplôme d'Ingénieur from INSA in 1991, and a Ph.D. from University of Rennes in 1995, both in Computer Science.
Prior to joining ANU, I held appointments with INRIA (1995-1997) and CSIRO (1997-2001). During 2003-2016, I had the opportunity to contribute to National ICT Australia (NICTA, now CSIRO Data61) in various roles, including that of Canberra Laboratory director (150 staff and PhD students).
Service to the community:
Service to the university:
ANU: See the projects I am offering and information about
competitive PhD scholarships.
Recent Awards
Research Projects
Australian Research Council:
ANU and NICTA:
Other:
The project is uniting computer scientists, philosophers, and social scientists in the pursuit of a more ethical future for AI and Machine Learning.
The focus of the project was the design and implementation the planning and optimisation algorithms required to operate smart energy grids reliably and economically.
The project has developed a spectrum of model-based methods for monitoring and diagnosing distributed dynamic systems. These methods rely on decentralised reasoning, symbolic representations, and model compilation.
DPOLP has made a number of significant contributions
to building tools for planning with time and uncertainty, and to
bridging the gap between Petri nets analysis, planning/heuristic
search in artificial intelligence, and statistical machine learning.
The project involves ANITI/Federal University of Toulouse,
KU Leuven, Saarland University, University of Bologna, Airbus,
Optit, and SciSports. It aims to obtain scalable, yet transparent,
robust and safe hybrid algorithmic solutions for planning \&
scheduling and demonstrate them on a number of use cases in
manufacturing, aircraft operations, sport management, waste
collection, and energy management.
The project involves UTAS (lead), ANU,
Tasnetworks, Powerlink Queensland, and Technische Universitaet
Berlin, and is also supported by the Australian-German Energy
Transition Hub. It developed approaches enabling consumer-owned DER
to participate in the Energy and Frequency Control Anciliary Service (FCAS)
market whilst complying with distribution network
constraints.
The project focused on planning
and scheduling under uncertainty for aircraft and satellite
applications. One of
the published
outcomes are approaches for 4D flight planning under whether
uncertainty, fuel consumption and other constraints.
CONSORT was a collaboration between
ANU, Tasmanian Networks, Reposit Power, The University of Sydney,
and The University of Tasmania. The project developed and deployed network-aware coordination
(NAC), a distributed approach for automatically coordinating consumer-owned PV-battery systems, so
that they simultaneously provide network support and value to their
owners. It trialled NAC on Bruny Island, Tasmania, to
reduce the use of diesel using peak events, and assessed the network's and households'
response to the technology as well as various payment structures.
See video
and the project's final reports
The project researches algorithms that support the
construction of risk-bounded mission plans, formulated as
constrained stochastic shortest path problems (CSSPs) with probabilistic
temporal logic objectives. The outcomes are the first heuristic search algorithms for
CSSPs (i-dual and i^2-dual), the first probability-aware heurstics for SSPs and CSSPs (hpom and hroc),
and extensions of these to multi-objective probabilistic LTL constraints.
With additional support from NICTA, the project built SmartGridToolbox, an extensible and interfaceable smart grid and microgrid simulator, suitable to generate and evaluate a range of control, optimisation and demand management strategies.
Professional Activities
Students
Prospective Students:
If you are contemplating doing your summer scholarship project,
honours project, Masters project, or a PhD in Artificial Intelligence,
please contact me, before applying via the usual channels.
Current Students:
Past Students:
Papers
This document is maintained by
Correction: The complexity result in Theorem 4 is incorrect. The
incorrectnes is based on an oversight in the use of a Tseitin-style
CNF transformation when progressing LTL formulas to the next
state. A correct progression-based algorithm, of higher complexity
though, employs standard CNF instead. The soundness theorem (Theorem
5), and completeness theorem (Theorem 6) still apply with that
correction. Moreover, and importantly, it is such a correct version
that we had implemented and used in our experiments. The
experimental results, hence, are not affected by the incorrectnes,
and neither are the other results.
[ps.gz pdf]
Extended (but older) version:Technical Report CMIS 98/140, CSIRO Math.
& Info. Sc., Canberra (Australia), July 1998.
[ps.gz] © CSIRO
More about the topic: see the
Blocks World page
Short version: 2nd European Workshop on Planning (EWSP-93),
pages 292-305, Vasdena (Sweden), IOS Press, December 1993.
[ps.gz] © IOS Press
Alternative version somewhat extended, with a different example:
[ps.gz] © the authors
Very short version: 8th International Symposium on Methodologies for
Intelligent Systems (ISMIS-94),
pages 305-314, Charlotte (NC), LNAI 869, October 1994.
[ps.gz] © Springer
Last Modified December 2024
