ORAHS 2015

HEC Montréal, July 19 - 24, 2015

/system/images/000/000/306/logos3_default.png

ORAHS2015

HEC Montréal, July 19 — August 31, 2015

Schedule Authors My Schedule
Cal add eabad1550a3cf3ed9646c36511a21a854fcb401e3247c61aefa77286b00fe402

ThA Tutorial 2

Jul 23, 2015 09:00 AM – 10:00 AM

Location: BDC

Chaired by Angel Ruiz

1 Presentation

  • Cal add eabad1550a3cf3ed9646c36511a21a854fcb401e3247c61aefa77286b00fe402
    09:00 AM - 10:00 AM

    Addressing uncertainty in health care with the cardinality-constrained approach: a trade-off between accuracy, computational effort, and interaction with clinicians

    • Ettore Lanzarone, presenter, CNR-IMATI

    Uncertainty is a fundamental aspect of several health care optimization problems, which cannot be neglected due to the significant impact it may have both on quality and feasibility of the problem
    solution. Indeed, high uncertainty is always related to patients’ conditions and demands, and the
    solutions should guarantee a good quality of the service over a usually wide number of possible future
    realizations.
    Different approaches have been proposed and applied in the literature to deal with uncertainty in health care problems, which can be mainly classified into stochastic programming, distributionally robust optimization, and robust optimization. Within the robust optimization approaches, the cardinality-constrained approach, introduced by Bertsimas and Sim about 10 years ago, represents a powerful tool that allows a trade-off between the level of robustness and the computational cost of the solution. Moreover, an intuitive modeling of the uncertainty set, which can be understood and tuned by clinicians and planners without any background in operations research, is another advantage of this approach. However, despite its potentialities, this approach has been only marginally applied in the health care sector.
    Briefly, the cardinality-constrained approach assumes that all of the uncertain parameters belong to an interval around a nominal value, and concentrates the variability of the problem by assuming that in each constraint only a limited number of parameters deviate from the nominal to the maximum value.
    The tutorial deals with the discussion of the robustness concept in health care optimization and with a detailed description of the cardinality-constrained approach: the methodology is presented and some
    examples of application to health care are given, e.g., to the assignment problem. Finally, an extension of the approach is outlined, and some remarks and drawbacks to take into account while implementing the approach in the practice are discussed.

Back