SAMU project
Back in 2020, I started with Anton Kleywegt from Georgia Institute of Technology and Victor Hugo Nascimento a project on the management of a fleet of ambulances that has since then become the most challenging project of applied mathematics and software engineering I have been involved so far. Originally, our goal was to design optimized allocation strategies of ambulances to emergency calls, in collaboration with Rio de Janeiro Emergency Health Service (EHS). This study gave birth to our first paper on the subject which is https://lnkd.in/d7Yb49rT
With my friend and school mate Marc Dutoo and the two software engineers Thais Viana and Victor Correa, we then developped the SOSZen software suite released in 2023 for EHS.
Our journey on the applied mathematics side was fed by our discussions along the way with Rio de Janeiro EHS, yielding studies on the calibration of statistical models for emergency calls and on allocation strategies of ambulances to emergency calls.
We developed LASPATED, see https://lnkd.in/drnC5NuR and https://lnkd.in/dAqYmTrK, a software for the analysis of spatiotemporal data and for the calibration of models for such data. LASPATED can in particular be used to calibrate models for emergency calls.
In [Heuristics](https://lnkd.in/d87qk7Vq), we proposed new heuristics for the ambulance selection problem and the ambulance reassignment problem. In the ambulance selection problem, one has to decide what to do when a call arrives: place this call in a queue of calls or send an ambulance immediately to that call (in this case, which ambulance?). For the ambulance reassignment problem, we have to decide what to do when an ambulance finishes service: send that ambulance to a call in queue or send it to a cleaning or parking base. Our heuristics are used in both rollout and nonrollout modes and outperform all 5 competing heuristics from the literature.
The following webpage is the academic window of our project.
View Project