{"id":{"repo_id":"toronto-retro","oai_identifier":"oai:utoronto.scholaris.ca:1807/91853"},"canonical_url":"https://search.dev.ndltd.org/etd/toronto-retro/oai:utoronto.scholaris.ca:1807/91853","repository":{"repo_id":"toronto-retro","name":"University of Toronto","base_url":"https://utoronto.scholaris.ca/server/oai/request"},"display":{"title":"Emergency Medical Services Response Optimization","abstract":"Time-sensitive medical emergencies are responsible for one-third of all deaths worldwide and similarly affect both developed and developing countries. Emergency medical services (EMS) provide rapid out-of-hospital treatment for time-sensitive medical emergencies. In this thesis, we combine optimization and machine learning to solve challenging EMS response problems in two diverse settings: Dhaka, Bangladesh - an urban center in a developing country, and Toronto, Canada - an urban center in a developed country. These settings are unified by uncertainty and stochasticity leading to new robust and chance-constrained optimization problems. In both cities, we employ machine learning to integrate real data with our optimization models. The second chapter develops a unified framework for emergency response optimization under travel time (edge-length) and demand (node-weight) uncertainty that is suitable for developing urban centers. We traveled to Dhaka to conduct field research resulting in the collection of two unique datasets that we leverage to estimate demand for EMS and to predict travel times in the road network. We carefully integrate our predictions and robust optimization model to develop an efficient solution algorithm for large-scale problems. We use our framework to provide an in-depth investigation into four key policy-related questions. The third and fourth chapters focus on improving EMS response for out-of-hospital cardiac arrest (OHCA). Chapter 3 employs a simplified location-queuing framework to quantify the potential benefit from using drones to deliver automated external defibrillators (AEDs) to OHCAs. We demonstrate, using data from 50,000 historical OHCAs covering 26,000 square kilometers around Toronto, that a drone network has the potential to significantly reduce AED delivery time. Chapter 4 develops a two-stage machine learning approach to simulate cardiac arrest incidents and an integrated location-queuing model tailored to the problem of drone-delivered AEDs. Our model combines the p-median framework with an explicit M/M/d queue to determine the minimum number of drones required to meet a pre-specified response time goal (average or 90th percentile), while guaranteeing that a sufficient number of drones are located at each base. We develop a novel reformulation technique that exploits the baseline (EMS) response times, allowing us to optimally solve large-scale instances and provide policy insights.","abstract_html":"Time-sensitive medical emergencies are responsible for one-third of all deaths worldwide and similarly affect both developed and developing countries. Emergency medical services (EMS) provide rapid out-of-hospital treatment for time-sensitive medical emergencies. In this thesis, we combine optimization and machine learning to solve challenging EMS response problems in two diverse settings: Dhaka, Bangladesh - an urban center in a developing country, and Toronto, Canada - an urban center in a developed country. These settings are unified by uncertainty and stochasticity leading to new robust and chance-constrained optimization problems. In both cities, we employ machine learning to integrate real data with our optimization models. The second chapter develops a unified framework for emergency response optimization under travel time (edge-length) and demand (node-weight) uncertainty that is suitable for developing urban centers. We traveled to Dhaka to conduct field research resulting in the collection of two unique datasets that we leverage to estimate demand for EMS and to predict travel times in the road network. We carefully integrate our predictions and robust optimization model to develop an efficient solution algorithm for large-scale problems. We use our framework to provide an in-depth investigation into four key policy-related questions. The third and fourth chapters focus on improving EMS response for out-of-hospital cardiac arrest (OHCA). Chapter 3 employs a simplified location-queuing framework to quantify the potential benefit from using drones to deliver automated external defibrillators (AEDs) to OHCAs. We demonstrate, using data from 50,000 historical OHCAs covering 26,000 square kilometers around Toronto, that a drone network has the potential to significantly reduce AED delivery time. Chapter 4 develops a two-stage machine learning approach to simulate cardiac arrest incidents and an integrated location-queuing model tailored to the problem of drone-delivered AEDs. Our model combines the p-median framework with an explicit M/M/d queue to determine the minimum number of drones required to meet a pre-specified response time goal (average or 90th percentile), while guaranteeing that a sufficient number of drones are located at each base. We develop a novel reformulation technique that exploits the baseline (EMS) response times, allowing us to optimally solve large-scale instances and provide policy insights.","abstract_has_math":false,"creators":["Boutilier, Justin James"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":"Mechanical and Industrial Engineering","school":null,"contributors":[],"advisors":["Chan, Timothy CY"],"committee_chairs":[],"committee_members":[],"year":2018,"date_issued":"2018-11","date_published":"2018-11","updated_at":"2026-07-27T21:28:16Z","subjects":["Drones","Emergency medical services","Facility location","Global health","Machine learning","Optimization"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/1807/91853","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Chan, Timothy CY"]},{"key":"dc:contributor.department","label":"Department","values":["Mechanical and Industrial Engineering"]},{"key":"dc:creator","label":"Author","values":["Boutilier, Justin James"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2018-11"]},{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2018-11-17T00:01:21Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2018-11-17T00:01:21Z"]},{"key":"dc:date.issued","label":"Date","values":["2018-11"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Drones","Emergency medical services","Facility location","Global health","Machine learning","Optimization"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["http://hdl.handle.net/1807/91853"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Time-sensitive medical emergencies are responsible for one-third of all deaths worldwide and similarly affect both developed and developing countries. Emergency medical services (EMS) provide rapid out-of-hospital treatment for time-sensitive medical emergencies. In this thesis, we combine optimization and machine learning to solve challenging EMS response problems in two diverse settings: Dhaka, Bangladesh - an urban center in a developing country, and Toronto, Canada - an urban center in a developed country. These settings are unified by uncertainty and stochasticity leading to new robust and chance-constrained optimization problems. In both cities, we employ machine learning to integrate real data with our optimization models. The second chapter develops a unified framework for emergency response optimization under travel time (edge-length) and demand (node-weight) uncertainty that is suitable for developing urban centers. We traveled to Dhaka to conduct field research resulting in the collection of two unique datasets that we leverage to estimate demand for EMS and to predict travel times in the road network. We carefully integrate our predictions and robust optimization model to develop an efficient solution algorithm for large-scale problems. We use our framework to provide an in-depth investigation into four key policy-related questions. The third and fourth chapters focus on improving EMS response for out-of-hospital cardiac arrest (OHCA). Chapter 3 employs a simplified location-queuing framework to quantify the potential benefit from using drones to deliver automated external defibrillators (AEDs) to OHCAs. We demonstrate, using data from 50,000 historical OHCAs covering 26,000 square kilometers around Toronto, that a drone network has the potential to significantly reduce AED delivery time. Chapter 4 develops a two-stage machine learning approach to simulate cardiac arrest incidents and an integrated location-queuing model tailored to the problem of drone-delivered AEDs. Our model combines the p-median framework with an explicit M/M/d queue to determine the minimum number of drones required to meet a pre-specified response time goal (average or 90th percentile), while guaranteeing that a sufficient number of drones are located at each base. We develop a novel reformulation technique that exploits the baseline (EMS) response times, allowing us to optimally solve large-scale instances and provide policy insights."]},{"key":"dc:description.degree","label":"Dc Description Degree","values":["Ph.D."]},{"key":"dc:title","label":"Title","values":["Emergency Medical Services Response Optimization"]}]}],"canonical_facts":{"dc:contributor.advisor":["Chan, Timothy CY"],"dc:contributor.department":["Mechanical and Industrial Engineering"],"dc:creator":["Boutilier, Justin James"],"dc:date":["2018-11"],"dc:date.accessioned":["2018-11-17T00:01:21Z"],"dc:date.available":["2018-11-17T00:01:21Z"],"dc:date.issued":["2018-11"],"dc:description.abstract":["Time-sensitive medical emergencies are responsible for one-third of all deaths worldwide and similarly affect both developed and developing countries. Emergency medical services (EMS) provide rapid out-of-hospital treatment for time-sensitive medical emergencies. In this thesis, we combine optimization and machine learning to solve challenging EMS response problems in two diverse settings: Dhaka, Bangladesh - an urban center in a developing country, and Toronto, Canada - an urban center in a developed country. These settings are unified by uncertainty and stochasticity leading to new robust and chance-constrained optimization problems. In both cities, we employ machine learning to integrate real data with our optimization models. The second chapter develops a unified framework for emergency response optimization under travel time (edge-length) and demand (node-weight) uncertainty that is suitable for developing urban centers. We traveled to Dhaka to conduct field research resulting in the collection of two unique datasets that we leverage to estimate demand for EMS and to predict travel times in the road network. We carefully integrate our predictions and robust optimization model to develop an efficient solution algorithm for large-scale problems. We use our framework to provide an in-depth investigation into four key policy-related questions. The third and fourth chapters focus on improving EMS response for out-of-hospital cardiac arrest (OHCA). Chapter 3 employs a simplified location-queuing framework to quantify the potential benefit from using drones to deliver automated external defibrillators (AEDs) to OHCAs. We demonstrate, using data from 50,000 historical OHCAs covering 26,000 square kilometers around Toronto, that a drone network has the potential to significantly reduce AED delivery time. Chapter 4 develops a two-stage machine learning approach to simulate cardiac arrest incidents and an integrated location-queuing model tailored to the problem of drone-delivered AEDs. Our model combines the p-median framework with an explicit M/M/d queue to determine the minimum number of drones required to meet a pre-specified response time goal (average or 90th percentile), while guaranteeing that a sufficient number of drones are located at each base. We develop a novel reformulation technique that exploits the baseline (EMS) response times, allowing us to optimally solve large-scale instances and provide policy insights."],"dc:description.degree":["Ph.D."],"dc:identifier.uri":["http://hdl.handle.net/1807/91853"],"dc:subject":["Drones","Emergency medical services","Facility location","Global health","Machine learning","Optimization"],"dc:title":["Emergency Medical Services Response Optimization"],"dc:type":["Thesis"]},"updated_at":"2026-07-27T21:28:16Z"}