2022 Optimization Days

HEC Montréal, Québec, Canada, 16 — 18 May 2022

Schedule Authors My Schedule

TA4 - Multiobjective optimization

May 17, 2022 10:30 AM – 12:10 PM

Location: Dutailier International (green)

Chaired by Maria Camila Gomez Lopez

3 Presentations

  • 10:30 AM - 10:55 AM

    A Multiobjective Approach for Sector Duration Optimization in Stereotactic Radiosurgery Treatment Planning

    • Oylum Şeker, presenter, University of Toronto
    • Mucahit Cevik, Ryerson University
    • Merve Bodur, University of Toronto
    • Young Lee-Bartlett, Elekta Instrument AB
    • Mark Ruschin, Sunnybrook Health Sciences Centre

    Sector duration optimization (SDO) is a problem arising in treatment planning for stereotactic radiosurgery on Gamma Knife. We present a multiobjective linear programming model for SDO to generate a diverse collection of solutions so that clinicians can select the most appropriate treatment. We develop a generic two-phase solution strategy based on the epsilon-constraint method for solving multiobjective optimization models, which aims to systematically increase the number of high-quality solutions obtained, instead of conducting a traditional uniform search. We also propose an alternative version of our two-phase strategy, which makes use of machine learning tools to reduce the computational effort. In our computational study on eight previously treated real test cases, a significant portion of the solutions from our two-phase strategy outperformed clinical results and those from a single-objective model from the literature. Our experiments illustrate the usefulness of machine learning strategies to reduce the overall run times nearly by half while maintaining or besting the clinical practice.

  • 10:55 AM - 11:20 AM

    Resource allocation in collaborative reforestation value chains: Multi objective optimization with case study

    • Mahtabalsadat Mousavijad, presenter,
    • Luc LeBel,
    • Lehoux Nadia,
    • Caroline Cloutier,

    Seed allocation is a critical decision in the reforestation value chain. If the most suitable seeds are selected in accordance with the parameters of reforestation sites, reforestation becomes more successful, and this has positive effects on the forest supply chain. This research proposes five optimisation models testing combinations of three objectives for seed allocation in a collaborative reforestation value chain. These models account for various biological parameters of seed and ecological parameters of reforestation sites. The main objective for all models is to find the most compatible seeds for all reforestation sites. The other objectives are minimising the allocation of rare seed lots for balancing seed center inventory and minimising the diversity of allocated seed lots. The Quebec province is used as a case study.

  • 11:20 AM - 11:45 AM

    Multicriteria robust optimization model based on interval analysis for IMRT treatment plans

    • Maria Camila Gomez Lopez, presenter, UQAM, EAFIT University
    • Janosch Ortmann, UQAM
    • Ana María Anaya Arenas, ESG - UQAM
    • Maria Eugenia Puerta Yepes, EAFIT University
    • Juan Carlos Rivera Agudelo, EAFIT University

    Radiation therapy is a form of cancer treatment that uses ionizing radiation to treat the disease. One of the most advanced forms of treatment is called Intensity modulated radiotherapy (IMRT) which is a technique that allows better dose conformation to the tumor while reducing the dose in healthy tissue. Treatment plans are designed using optimization techniques and computational methods that allow the radiation dose to be calculated. To minimize the impact of errors in patient positioning that can potentially have a negative impact on the efficacy and safety of the treatment, robust optimization methods have been proposed. In general, a way to model uncertainty is by representing it as interval type structure. In this talk, I will discuss the theoretical framework of interval analysis, robust optimization and multicriteria optimization techniques and explain how it applies to the problem of inverse planning of radiotherapy treatments. The outcome are IMRT treatment plans that account for uncertainty and at the same time, help specialists make decisions regarding the best possible treatment plan for their patients.