An algorithm makes it possible to organize humanitarian aid after a natural disaster | technology

The earthquakes that devastated Nepal between April and May 2015 killed more than 9,000 people, injured 22,000 and affected nearly a million households. The survivors and their country were left in ruins, and in the most affected areas, about a million people were left without drinking water, because it only reached them through a network of public fountains, destroyed by earthquakes. The solution to this problem originated in Canada, where a group of researchers from HEC Montreal created a mathematical algorithm to draw a complete water supply plan without ever visiting the site. The tragedy is now being repeated in Turkey and Syria, and this project can be applied, with some modifications, to the distribution of humanitarian aid after the two earthquakes, which have already left more than 11 thousand dead in both countries.

Pompeu Fabra University postdoctoral researcher and co-author of the project, Jessica Rodríguez-Pereira, details the possibility of making “computational modifications to the model” and, using geographic data from the area, determining the most appropriate locations for, for example, doing the distribution of food and bottled water in bottles or for the preparation of medical services. “You won’t be able to open many positions and you’ll want people to be close, so [el programa] You can strike that balance to identify help points,” he explains. Together with Selina Silvestri, Ph.D. in mathematics and computer science, this expert received the award from the Statistics and Operations Research Association and the BBVA Foundation on January 18 for this work, as the best contribution applied to a real-world problem. social impact.

Their goal was to create the best points for building new community water sources in Nepal’s Dolakha region, close to the Himalayas, taking into account the geography and needs of the community. But some factors made this puzzle more complicated: The affected areas were located in a mountainous, mostly rural area, with few resources and a vulnerable population. In addition, the network that will connect the springs to the sources must be gravity, that is, without pumping water.

The ideal, explains Celine Silvestri, is for everyone to have a source at home, but this “wasn’t possible”. Therefore, as few water stations as possible should be built, according to a standard set by the authorities: people should not walk more than 250 meters horizontally or 80 meters vertically to the nearest source.

The team used satellite imagery, provided by a team from the University of Salzburg in Austria, to learn about the mountainous terrain, the location of the springs, and the locations of the residents. Jessica Rodriguez Pereira details the process of translating images into mathematical language. “We have a map, but we have turned it into a grid, where each cell is defined by coordinates. So we use these numbers and build a graph with points and connections between the points. Then we forget the map and save the data we got, ”he told EL PAÍS, before the award ceremony at the BBVA Foundation.

Rodríguez-Pereira asserts that after a disaster, similar tasks are usually carried out based on the experience of one or more people who know the terrain and draw a reconstruction plan “by hand”. “We don’t have a handcrafted solution for comparison. But in other applications we know that there are general improvements when using mathematics,” says the professor at the Barcelona School of Economics.

The algorithm developed by the team is somewhat similar to that used by companies like Amazon to organize the logistics of package distribution. Behind the speed of delivery is a mathematical formula and code that processes various data very quickly: the starting point, the destination point, and the possible paths. “They have several people in different places to deliver the packages. If the delivery person decides to deliver them using alphabetical order, this method might work, but it’s not the smartest way. The best thing to do is look where they are,” says Celine Silvestri, who traveled from Italy to Madrid to receive the award. And try to find the shortest way to visit them all.

It is a very complex task for a human being to undertake, as it requires the study of very many groups. On the other hand, mathematical intelligence and computers can easily solve it. Silvestri, who is currently working for data analytics firm FICO, believes that the beauty of mathematical optimization is the ability to translate a problem into a numerical model that, of all possible solutions, finds the best. This kind of approach is appropriate when you have to save money and time and meet people’s needs. It was very nice and interesting. And also a challenge for many reasons,” he stresses.

The solution was implemented in Nepal nearly four years after the earthquakes, and by that time local authorities had already rebuilt, albeit improvised, new water sources for people to survive. “We’re presenting a whole network, from scratch, as if nothing was there. But they needed to do something over time, so the end result is probably a mixed bag. We’re introducing a scientific tool to analyze what they’ve already done and what they’re still doing. need to do,” Silvestri wonders.

The mathematical formula used in the project is available in the article published in the journal Computer and Operations Research And the team is ready to cooperate with future applications, as in the case of Turkey and Syria. An NGO or a government organization can contact us and we can run the software. Or if there’s a tech team with knowledge of operational research and algorithms, they can also use the code directly,” Jessica Rodriguez Pereira says by phone this Tuesday.

to another network of Gravity water distribution, the only thing we have to change is the data. If it is another system, we have to modify the part of the algorithm that determines the distribution and the rest will be preserved. “It is possible to adapt it,” says the researcher, and continues: “It is the first time we have so much visibility and celebration. Young people in our field. Generally, you see your teachers there, not you, ”concludes Rodríguez Pereira, who received the award with emotion next to his colleague.

Scientists Gilbert Laporte and Marie-Yves Rancourt, both of HEC University Montreal, also received the prize, which was awarded in the amount of 6,000 euros. The work was developed between 2017 and 2019, a period when Rodíguez-Pereira and Silvestri were doing postdoctoral studies at the Canadian institution.

You can follow country technology in Facebook And Twitter Or sign up here to receive The weekly newsletter.