4 Presentations

• 
  02:00 PM - 02:25 PM

  Optimizing Operations at the CHUSJ Pediatric ICU: A Decision-Support Dashboard

  • Nabila Remli, presenter, ivado labs
  • Carl Perreault-Lafleur, ivado labs
  • Steven Lamontagne, ivado labs
  • Qing Liu, ivado labs
  • Guy Desaulniers, ivado labs
  • Guillaume Rabusseau, ivado labs
  • Louis-Martin Rousseau, ivado labs

  At the Sainte-Justine Pediatric Intensive Care Unit, we developed a decision-support dashboard to assist Assistant Chief Nurses (AICs) in managing patient care logistics. This tool provides a continuous overview of patients, their clinical and nursing needs, current room allocations, and nurse assignments, along with all essential service requirements. At the core of the system is an integer optimization model that streamlines the patient-to-room and nurse-to-patient assignment process at the beginning of each shift, ensuring alignment with patient needs, nurse specializations, room constraints, and AIC preferences. To support longer-term planning, we also developed a predictive model that forecasts the total nursing workload up to seven days in advance, enabling better staffing decisions and reducing the mandatory overtime.

• 
  02:25 PM - 02:50 PM

  Scheduling Radiation Therapy Personnel: A Lexicographic Mixed-Integer Programming Approach

  • Aldair Alvarez Diaz, presenter, Ivado Labs
  • Mathieu Sylvestre, Ivado Labs
  • Caroline Rocha, Ivado Labs
  • Guy Desaulniers, GERAD - Polytechnique Montréal
  • Guillaume Rabusseau, Université de Montréal
  • Louis-Martin Rousseau, Polytechnique Montréal

  Hospitals face many complexities in managing radiation therapists’ schedules, with the need to cover many work areas with qualified staff while balancing fairness, preferences, costs, and regulatory constraints. To address these complexities, we developed a scheduling system to assign radiation therapists to shifts across different frameworks distinguished by the immediacy of the decision to take and the length of their planning horizon. The proposed solution is based on a mixed-integer programming (MIP) model with lexicographic multi-objective optimization, allowing the prioritization of different goals depending on the planning horizon. The key technical challenge lies in encoding a diverse and sometimes conflicting set of objectives and constraints into a unified model while ensuring tractability and robustness in real-world deployment.

• 
  02:50 PM - 03:15 PM

  Intelligent Distribution of Supplies at the Jewish General Hospital

  • Louis Popovic, presenter, Ivado Labs
  • Mathieu Sylvestre, Ivado Labs
  • Yossiri Adulyasak, HEC Montréal
  • Jorge Mendoza, HEC Montréal

  Hospitals rely on timely and accurate replenishment of medical supplies to ensure uninterrupted patient care. At the Jewish General Hospital, many medical supplies are stocked in a centralized warehouse from which care units are replenished. The replenishment of care units is challenging due to the number of products to manage, the variability of demand, and the limited visibility into inventory levels within the care units. We devised a solution to optimize the replenishment of care units from the warehouse, aiming to reduce the risk of stockouts while minimizing logistical efforts. First, a demand forecast at the care unit–product level is generated daily, using historical requisitions and inventory snapshots. Using this predicted demand, a set of viable replenishment options is generated for each product and care unit, and the optimal plan is selected based on multiple competing objectives. This multi-objective optimization framework enables anticipatory replenishments that improve logistical efficiency. A custom user interface surfaced through tablets allows users to act on recommendations and validate or adjust inventory estimates feeding the recommendations, enabling the system to improve over time with continued use.

• 
  03:15 PM - 03:40 PM

  From Data to Decisions with Generative AI and Optimization Technology

  • Carlos Zetina, presenter, FICO
  • Dinakar Gade, The Ohio State University

  Your team of analytics experts has built state-of-the-art tools to drive business impact at your company. However, the tools remain cumbersome to use and are not accessible by business domain experts and users. In this tutorial, you’ll learn how FICO’s Xpress Optimization capabilities make it possible to embed your analytic and optimization models within business user-friendly applications. See how to supercharge your models with simulation, fast and robust optimization, reporting, what-if analysis, and agile extensibility for your ever-changing business. We will demonstrate how LLMs can help develop models and configure the solution as well as efficiently guide the business user.