Nairobi Ambulance Challenge

Our second competition team has started competing in the Nairobi Ambulance Challenge. This competition is hosted on the platform Zindi and is about finding the most optimal way of locating ambulance in the city of Nairobi to minimize the driving time to accidents.

Road traffic accidents are one of the highest causes of death worldwide. Emergency response is one of the critical aspects to save lives, where the delay of an ambulance arriving at a car crash can make the difference between death and life. This challenge aims to reduce this response time using AI. In this competition, thousands of locations of traffic accidents in Nairobi, Kenya, are provided. The goal is to place six ambulances in windows of 3 hours to minimize the total response time to all crashes. Car crashes are distributed differently depending on the weather, the time of the day, or the month.

The problem can be tackled as two different sub-problems. First, it is necessary to predict where future crashes will occur. Then, the six ambulances should be placed in an optimal way to minimize the arrival time to each of the crashes.

To solve the first problem, we decided to cluster the crashes with similar characteristics, such as the time of the day, the raining conditions, or the temperature, using K-means algorithms. As an example, you can see how a 2D K-means clustering works in the figure below. The second problem was solved using genetic algorithms. These algorithms are inspired by the process of natural selection, where each individual tries to perform as well as possible in the task, and the fittest are selected. If you want to know more details about how we did in this competition, join the Epoch Engineering channel!