3 Presentations

• 
  03:00 PM - 03:30 PM

  Mathematical programming models for radiotherapy scheduling with time windows

  • Bruno Vieira, presenter,
  • Derya Demirtas, University of Twente
  • Jeroen B. van de Kamer,
  • Erwin W. Hans, University of Twente
  • Nadia Lahrichi, Polytechnique Montréal
  • Louis-Martin Rousseau, Polytechnique Montréal
  • Wim H. van Harten,

  In external-beam radiotherapy (RT), high-energy radiation beams are delivered by a linear accelerator in a series of irradiation sessions undertaken over multiple days. In this work, we consider the problem of scheduling and sequencing RT sessions considering time window preferences given by patients for the starting time of their appointments. Most studies in the literature focus on assigning patients to linacs and days, neglecting the sequencing component, and existing sequencing algorithms are only able to solve the problem using approximation methods due to the intractability of the formulated models. We propose two mixed-integer linear programming formulations, and test them using data from a large Dutch RT center. Results show that the problem can be solved in reasonable computation time for real-world size instances using either model.

• 
  03:30 PM - 04:00 PM

  Pattern-based online algorithms for a general patient-centred radiotherapy scheduling problem

  • Roberto Aringhieri, presenter, Dipartimento di Informatica, Università degli Studi di Torino
  • Davide Duma, Dipartimento di Informatica, Università degli Studi di Torino
  • Giuseppe Squillace, Università degli Studi di Torino

  A radiation therapy, say also radiotherapy, is delivered by a linear accelerator or linac, which is a special device whose main function is to concentrate in beams and accelerate the emission of subatomic particles. A radiotherapy treatment consists in a given number of radiation sessions, one for each (working) day, which should start before a given release date. Patients are usually classified into classes of urgency having different deadlines and number of sessions. Waiting time is the main critical issues in the management of a radiotherapy health system. After deriving a general problem statement from the case studies reported in the literature, we present three online optimisation algorithms that try to exploit the particular structure of the solution, and we compare their results with two baseline online algorithms.

• 
  04:00 PM - 04:30 PM

  Multi-level heuristic to optimize the chemotherapy production and delivery

  • Alexis Robbes, presenter, Université de Tours
  • Yannick Kergosien, Université François Rabelais Tours
  • Jean-Charles Billaut, Université de Tours

  The bio pharmaceutical unit of Oncology Clinic (UBCO) of the hospital Bretonneau of Tours (France) produces between 100 and 300 injections per days for the three hospitals of Tours. The production of chemotherapy drugs consists of two steps: a sterilization step and a preparation step performed by pharmacists. The production can be modeling as a hybrid flow shop scheduling problem variation. Once the drugs are made, they must be delivered to the patient at a given due date thanks to delivery men. The delivery problem is a multi-trip vehicle routing problem variation. We proposed a multi-level heuristic to solve both production and delivery problems. Computational experiments are conducted on real-based instances to compare multiple settings and to evaluate the efficiency of the proposed approach.