Supply chain optimization increases UN World Food Programme’s impact
Thanks to predictive analytics, network design and optimized transport flows, many more people can be provided with food and emergency aid. This was stated by Hein Fleuren, Professor of Operations Research and co-founder of the Zero Hunger Lab, during his keynote at the event called ‘In het spoor van Ad van Goor’. ‘A breakthrough during our project at the UN World Food Programme came when we started looking at the food requirement through a supply chain lens.’
By Harm Beerens
‘By deploying data analysis and optimization tools, we can feed two million more people with the same budget.’ These words from 2021, by Amir Abdullah, the then-deputy executive director of the UN World Food Programme, are still music to the ears of Professor Hein Fleuren (pictured). It was a great compliment to his Zero Hunger Lab that he had founded 2 years earlier with his associate Perry Heijne, through which he had developed a supply chain optimization tool for distributing emergency aid. Since then, the model has saved international aid agencies a lot of money, as it did again recently with food supplies in Gaza. ‘And unlike in business, this money does not go to shareholders but to the people who need it most in this world,’ Fleuren says.
TNT Express bus timetable
‘How can logistics fight hunger?’. That was the title of his keynote at the knowledge and networking event organized annually in memory of Professor Ad van Goor, who died in 2020. Hein Fleuren is also a professor at Tilburg University, where he has been focused on operations research and predictive analytics since 2000. In partnership with others, he has won the Franz Edelman Award – the most prestigious recognition you can get as a scientist in this field – twice in his life. The first time was in 2012 for a supply chain network model that allowed TNT Express to distribute its parcels around the world as efficiently as possible. With this ‘bus timetable’, as he calls it, TNT was able to save €200 million in transport costs over four years.
Driving in convoys
Fleuren begins his talk in Amersfoort with some sad facts about the very unequal distribution of food in the world. Every day, 811 million people go to bed hungry and 3 billion cannot afford a healthy meal. In Europe, in contrast, there is a surplus of food; per person, 127kg are thrown away or wasted every year. ‘For the UN World Food Programme, we initially developed a model that allows you to better align food supply and demand. This is an incredibly complicated puzzle with lots of nodes, but therefore is wonderful for operations research fanatics to work on. You have to take into account restrictions that we don’t have at all in Europe, such as the need to drive in convoys. In countries where there is poverty, you can’t just drive a single truck on its own.’
Supply chain lens
By applying supply chain network design theory and cleverly combining transport flows, Fleuren and his team were able to present some significant efficiency improvements. However, the real breakthrough came when they started looking at the food requirement through a supply chain lens. Fleuren: ‘Until then, we’d assumed in our models that each person needs a certain amount of corn, grain, oils and other items every day. Until someone said: “Can’t we offer the same nutritional value with other products, ones that can be sourced closer to home?” For example, we transported red lentils from Canada to Chad, but the same nutrients are also available in broad beans that can be sourced in Turkey.’ By looking at the food requirement through more of a supply chain lens, huge savings could be made on transport distances.
Predictive analytics
Disaster relief is another United Nations project the Zero Hunger Lab contributed to. ‘During earthquakes, floods or other disasters, programmes are launched to meet acute needs for food, clothing and medicine. With the Red Cross, we are doing a study that shows that better forecasting allows much of this relief provision to be started earlier, i.e. before the disaster actually occurs. We have also calculated that, with these anticipatory measures, the delivery itself can be done much faster and more efficiently. Every dollar you invest in the phase prior to an approaching disaster saves you $7 in relief provision afterwards.’
Many people are still getting used to the fact that predictive analytics and taking action in advance can make disaster relief much more efficient, Fleuren observes. ‘The conclusion is that it would actually be much better to do post-disaster fundraising in advance. But people don’t want to give money for an impending famine, or for an earthquake that might happen one day. To change this, we are now developing a kind of model to show donors the effect of the way they give. Hopefully, this will allow agencies to raise money at the times when they can use it most optimally’.
