4 Presentations

• 
  10:30 AM - 10:55 AM

  Improving real-time ambulance redeployment decisions for emergency medical services

  • Sara Mesgari, presenter,
  • Jonathan Patrick, Telfer School of Management, University of Ottawa
  • Antoine Saure, University of Ottawa
  • Mohi Sedighi, Telfer school of management

  Emergency medical services (EMS) play a critical role in responding to medical emergencies and transporting patients between healthcare centers. Currently, EMS face challenges such as a shortage of staff and ambulances, lengthy response times, and delays in patient transport. To address these challenges, this research focuses on improving the allocation and relocation of EMS vehicles using mathematical programming, specifically dynamic programming. State, action, and objective components will be specified to minimize the number of unserved calls within a time threshold while considering the cost of vehicle movement. The proposed model incorporates uncertainty in the demand for services and the length of service and is being developed using Lanark County Paramedic Service data. Simulation will be used to evaluate its performance and compare it with that of the current dispatching approach. We expect the outcome of this research will provide a systematic decision-making approach to optimizing vehicle allocation and relocation and managerial insights regarding future capacity requirements to achieve specific service levels.
• 
  10:55 AM - 11:20 AM

  Deep learning-assisted appointment scheduling under uncertainty

  • Amirhossein Moosavi, presenter,
  • Ozturk Onur, Telfer School of Management, University of Ottawa
  • Jonathan Patrick, Telfer School of Management, University of Ottawa

  A well-designed appointment system has the potential to improve the utilization of expensive resources and reduce patient wait times in outpatient clinics. This study focuses on developing an appointment system for outpatient health centers that provide heterogeneous, multi-step care through multiple providers (similar to a flexible job-shop scheduling problem). Our objective is to optimize resource utilization, patient wait times, and staff overtime in the presence of uncertainties related to service duration, punctuality, and no-shows. To solve this problem, we use a learning-based algorithm that combines deep learning and a partial mixed integer programming algorithm. Specifically, we develop a predictive model using deep learning, a convolutional neural network, that accounts for uncertainties and estimates the associated costs. The predictive model is trained on a randomly generated dataset prior to the optimization process and subsequently is integrated into the objective function of a deterministic model. The methodology is evaluated by investigating a children's hospital in Canada that provides various plaster care (including examination, casting, or cast removal). We compare performance metrics of schedules with and without uncertainties and ensure their robustness via simulation.
• 
  11:20 AM - 11:45 AM

  Scheduling and routing of home care services with periodic visits and time windows

  • Peyman Varshoei, presenter, University of Ottawa, Telfer School of Management
  • Jonathan Patrick, Telfer School of Management, University of Ottawa
  • Onur Ozturk, Telfer School of Management, University of Ottawa

  In this study, we investigate a scheduling and routing problem for personal support workers (PSW). The problem involves scheduling and routing periodic visits for PSWs while adhering to various constraints such as clients' visit time preferences, PSWs' working hour limits, and breaks between and within shifts. The objective is to minimize travel times, maximize continuity of care for clients, minimize the number of outsourced PSWs, and minimize the violation of clients' preferred visit times. To tackle this NP-hard multi-objective combinatorial problem, we proposed a modified version of the nondominated sorting genetic algorithm II (NSGA-II). Our algorithm introduces novel route-creator procedures that offer greater exploration and diversity in the solution space. We evaluate the performance of our solution approach using actual data from a personal support provider company in Toronto for a two-week schedule. Our results demonstrate an improvement in continuity of care for clients, a reduction in the average idle time and travel time for each visit, and a significant decrease in the visits made by PSWs from other agencies. Overall, our proposed algorithm exhibits promising outcomes in terms of efficiency and effectiveness in creating a scheduling and routing plan in complex scheduling and routing problems.
• 
  11:45 AM - 12:10 PM

  An Approximate Dynamic Programming Approach to Network-Based Scheduling of Chemotherapy Treatment Sessions

  • Alejandro Cataldo, Universidad Católica de Chile
  • Pablo A. Rey, Universidad Tecnológica Metropolitana
  • Antoine Sauré, presenter, University of Ottawa
  • Arturo Wenzel, Pontificia Universidad Católica de Chile

  A solution approach is proposed for the interday problem of assigning chemotherapy sessions at a network of treatment centres with a view to increasing the efficiency of system-wide capacity use. This network-based scheduling procedure is subject to the condition that both the first and last sessions of a patient's treatment protocol are administered at the same centre the patient is referred to by their oncologist. All intermediate sessions may be administered at other centres. The problem is modelled as a Markov decision process which is then solved approximately using techniques of approximate dynamic programming. The benefits of the approach are evaluated and compared through simulation with the existing manual scheduling procedures at two treatment centres in Santiago, Chile. The results suggest that the approach would obtain a 20% reduction in operating costs for the whole system and cut existing first- session wait times by half. A key conclusion, however, is that a network-based scheduling procedure brings no real benefits if it is not implemented in conjunction with a proactive assignment policy like the one proposed.