CWI develops algorithms that shorten response time of ambulance

In emergency situations where every second counts, the timely presence of an ambulance can be a matter of live or death. In The Netherlands the norm for the response time of ambulances amounts 15 minutes for high-urgency incidents. In practice this norm is not always met however.

Publication date: 24-01-2017

In emergency situations where every second counts, the timely presence of an ambulance can be a matter of live or death. In The Netherlands the norm for the response time of ambulances amounts 15 minutes for high-urgency incidents. In practice this norm is not always met however.

Repositioning of vehicles

The response time of ambulances can be shortened by dividing the vehicles efficiently over the region. The location of the ambulance at the moment of the casualty decides for a large part whether or not the norm time will be met. PhD student Thije van Barneveld of CWI discusses several models and algorithms for the optimization of the distribution of ambulances throughout the region by executing so called proactive relocations. This means one or more vehicles will be repositioned temporarily. This is the essence of Dynamic Ambulance Management (DAM). Van Barneveld: “the great advantage of DAM is that by moving the vehicles real-time we can respond to future incidents much more flexible, whereby response times can be reduced and patients can be assisted much faster.“

Penalty heuristic

Van Barneveld has also focused on the disadvantages of DAM in his thesis. One of the criticisms regarding this method, is that ambulances have to relocate proactively quite often, causing the workload to rise. In this research new algorithms have been used to balance the decision between the number of relocations and the performance regarding response times. Penalty heuristic means that the supplier of the ambulance service will get a virtual penalty when the response times take longer. On the basis of input from experts from the field a penalty heuristic for a great number of scenarios has been examined and determined to which extent the number of relocations influence performance. By executing a small number of relocations performance already improves significantly. Frequent relocations can lead to a stabilization or even a deterioration of performance.

Online & Offline methods

The methods discussed can be divided in two categories: online and offline algorithms. The difference between both methods is the moment the calculations take place. When using online methods, this will be at the moment the decision to relocate the vehicle has to be made. For offline, the majority of the calculations are being made prior to the decision making moment. Van Barneveld’s research proves that by relocating ambulances proactively, response times can be shortened.

Defense: 25/01/2017 - 11:45 - 13:15

Vrije Universiteit, de Boelelaan 1105, Amsterdam

Promotor: Prof.dr. Rob van der Mei (CWI Stochastics/VU) and Prof.dr. Sandjai Bhulai (VU)

Thesis: ‘Relocation Algorithms for Emergency Medical Services’

The CWI spin-off Stokhos Emergency Mathematics provides software solutions which enable the proactive relocation of emergency services based on predictions of future demand.