Resilience

Natural and man-made disasters constitute a major threat to our economies, environments, and communities. Our research aims at providing unprecedented situation awareness and decision support for mitigating, responding, and recovering from disasters. These decisions involve multiple complex infrastructures, multiple agencies, and multiple stakeholders from state and federal agencies, to local councils and communities. Some of our techniques have been deployed in collaboration with our partner, the Los Alamos National Laboratories, to mitigate the effects of hurricanes on the coastal areas of the United States. Others were being developed in collaboration with some state emergency services in Australia.


Evacuations

Our research develops new planning and real-time scheduling algorithms for evacuations, capturing detailed models of the traffic network, the disaster thread, and human behaviour. The zone-based prescriptive algorithms make recommendations about when to evacuate residential zones, which route to follow and how to mobilize the warning resources, exploiting the information about the thread and minimising congestion in the traffic network. Our research also develops high-performance food simulation algorithms, exploiting parallel computing and GPUs, as well as advanced visualisation tools.

Power Restoration

Our research aims at minimising the size of a blackout after a major disaster. It combines the logistics aspects of power restoration (sending repair crews at the damaged sites) with the power system aspects (which parts of the network to restore first). This research requires sophisticated modelling of the power systems and associated computational methods, since it is not operating under normal conditions. The research considers both the steady states of the networks and transient stability.

Relief Distribution

Our research focuses on efficient optimisation algorithms to deliver relief during and after disasters. It considers the strategic, tactical, and operational scheduling of relief distribution, a computationally challenging last-mile logistic problems under uncertainty.


Recent Publications


2016

  • Optimal Flood Mitigation over Flood Propagation Approximations.
    Byron Tasseff, Russell Bent, and Pascal Van Hentenryck.
    Proceedings of the 14th International Conference on the Integration of Artificial Intelligence, and Operations Research Techniques in Constraint Programming (CP’AI’OR’16), Banff, Canada, May 2016.

  • Rapid assessment of disaster damage using social media activity.
    Yury Kryvasheyeu, Haohui Chen, Nick Obradovich, Esteban Moro, Pascal Van Hentenryck, James Fowler, Manuel Cebrian.
    Science Advances. 2(3), March, 11, 2016

  • Benders Decomposition for Prescriptive Evacuation Planning.
    Julia Romanski and Pascal Van Hentenryck
    Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence (AAAI-16), Phoenix, Arizona, February, 2016.

  • Optimizing Infrastructure Enhancements for Evacuation Planning.
    Kunal Kumar, Julia Romanski, and Pascal Van Hentenryck
    Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence (AAAI-16), Phoenix, Arizona, February, 2016.

  • Intelligent Habitat Restoration Under Uncertainty.
    Tommaso Urli, Jana Brotánková, Philip Kilby, and Pascal Van Hentenryck
    Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence (AAAI-16), Phoenix, Arizona, February, 2016.

  • A Conflict-Based Path-Generation Heuristic for Evacuation Planning.
    Victor Pillac, Pascal Van Hentenryck, and Caroline Even.
    Transportation Research Part B., 83, 136-150, January, 2016.

2015

2014

  • Power System Restoration Planning with Standing Phase Angle and Voltage Difference Constraints

    Terrence Mak, Carleton Coffrin, Pascal Van Hentenryck, Ian A. Hiskens, David Hill
    Proceedings of the Power Systems Computation Conference (PSCC), Wroclaw, Poland, pp. 8, August, 2014.
    This paper considers the restoration of a transmission system after a significant disruption such as a natural disaster. It considers the Restoration Order Problem (ROP) that jointly considers generator […]

  • Transmission System Restoration: Co-optimization of Repairs, Load Pickups, and Generation Dispatch

    Carleton Coffrin, Pascal Van Hentenryck
    Proceedings of the Power Systems Computation Conference, Wroclaw , Poland, pp. 7, August, 2014.
    This paper studies the restoration of a transmission system after a significant disruption (e.g., a natural disaster). It considers the co-optimization of repairs, load pickups, and generation dispatch to […]

  • NICTA Evacuation Planner: Actionable Evacuation Plans with Contraflows

    Caroline Even, Victor Pillac, Pascal Van Hentenryck
    Proceedings of the Prestigious Applications of Artificial Intelligence 2014 (PAIS2014), Prague, Czech Republic, pp. 1143-1148, August, 2014, 10.3233/978-1-61499-419-0-1143.
    Evacuations are a critical aspect of disaster management, and generally the first prevention measure to ensure the safety of persons under the threat of a forthcoming disaster. Designing evacuation […]

  • A path-generation matheuristic for large scale evacuation planning

    Victor Pillac, Pascal Van Hentenryck, Caroline Even
    Proceedings of the 9th International Workshop on Hybrid Metaheuristics, Hamburg, pp. 71-84, June, 2014, 10.1007/978-3-319-07644-7_6. (© Springer 2014)
    In this study we present a general matheuristic that decomposes the problem being solve in a master and subproblem. In contrast with the column generation technique, the proposed approach […]

2013

  • A Conflict-Based Path-Generation Heuristic for Evacuation Planning

    Victor Pillac, Pascal Van Hentenryck, Caroline Even
    Technical report, NICTA, Melbourne, September, 2013, 1833-9646-7393.
    Evacuation planning and scheduling is a critical aspect of disaster management and national security applications. This paper proposes a conflict-based path-generation approach for evacuation planning. Its key idea is […]

  • Computational Disaster Management

    Pascal Van Hentenryck
    Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI), Beijing, China, pp. -, August, 2013.
    The frequency and intensity of natural disasters has significantly increased over the past decades and this trend is predicted to continue. Natural disasters have dramatic impacts on human lives […]


Demonstrations