Researchers from Koç University and international collaborators have developed a new algorithm that enables faster and more equitable distribution of disaster relief supplies. By integrating fairness directly into logistics planning, the model reduces inequality in unmet demand by up to 34% without compromising delivery speed. The approach offers a practical tool for improving decision-making in real-world emergency response operations.

In the chaotic aftermath of a major earthquake, getting supplies to those in need becomes a race against time. But speed alone is not enough—fairness matters just as much. A study published in the European Journal of Operational Research by researchers from Koç University, Polytechnique Montréal, and the University of Vienna addresses one of the most pressing challenges in disaster logistics: how can limited relief supplies be delivered both quickly and fairly?

The Problem: Not Enough Supplies for Everyone
When the devastating 2023 Maraş earthquakes struck Türkiye, more than 2.5 million people required shelter. Emergency supplies had to reach hundreds of tent cities across eleven provinces, yet there simply was not enough to meet all needs, and moreover, poor road conditions delayed transportation. Some regions, such as Maraş and Gaziantep, received aid relatively quickly, while others, including Hatay and Adıyaman, faced significantly longer delays and higher unmet demand.

Speed vs. Fairness: A Delicate Balance
Led by Prof. Dr. Sibel Salman from Koç University’s College of Engineering, the research team developed a mathematical model and a solution algorithm to plan delivery truck routes and determine how much aid each shelter should receive.

One of the key innovations of the study is that it simultaneously optimizes routing and allocation decisions—determining not only where and when aid is delivered, but also how much each location should receive within a single integrated framework.

The model balances two competing objectives: minimizing total travel time to ensure rapid delivery and minimizing inequality in unmet demand across shelters so that no region is disproportionately disadvantaged.

Rather than simply measuring fairness, the researchers incorporate the Gini coefficient—widely used in economics to quantify inequality—directly into the optimization model, effectively turning fairness into a decision-making objective.

A Smarter Algorithm
Because the problem is highly complex, standard optimization tools are often too slow for real-world emergency scenarios. To address this, the team developed a specialized “branch-and-price” algorithm that breaks the problem into smaller components, solves them efficiently, and combines them into a high-quality overall plan.

This tailored approach significantly improves computational efficiency, enabling near-optimal solutions to be generated within time frames suitable for real disaster response operations.

Tested using real data from the 2011 Van earthquake and projected scenarios for Istanbul’s Kartal district, the algorithm outperformed commercial solvers and produced near-optimal solutions within practical time limits.

Improved Aid Delivery in Practice
The results are striking. The proposed approach reduces inequality in aid distribution by approximately 34% without compromising delivery speed.

The study also shows that when time constraints are either very loose or very tight, simply maximizing total coverage can yield good results. However, under more realistic, moderately constrained conditions, ignoring fairness leads to highly unequal outcomes. In such cases, a balanced strategy is not just preferable—it is essential.

This research offers a powerful tool for humanitarian organizations, enabling them to deliver aid rapidly without sacrificing fairness in future disaster response efforts.