Confident Data-Driven Decision Support
Funded by The European Union under The Marie Skłodowska-Curie Actions (MSCA) Doctoral Networks programme Grant Agreement No. 101227512
Organizations have a growing need to solve optimization problems to improve their competitiveness and sustainability, but traditional optimization methods do not incorporate data to enhance their performance, increase their robustness, or adjust decisions to ensure they are fair. To ensure this, and further compliance with principles of trustworthy AI, optimization methods should be enhanced by data analytics and machine learning (ML) approaches, resulting in data-driven optimization (DDO) methods. However, the expertise needed to create and apply DDO methods to real-world problems is severely lacking, as DDO experts must be trained in highly advanced methods in both ML and operations research (OR). Compounding the challenges facing organizations is the fact that current DDO methods generally do not confirm to the principles of trustworthy AI, such as fairness, robustness, and explainability/interpretability. The Global Partnership for AI (GPAI), which includes the EU, calls for strengthening the trustworthy AI “ecosystem” to “accelerate change”, but in the area of DDO, the methods necessary for the ecosystem do not yet exist. Creating trustworthy optimization methods requires the cross fertilization of ML and OR, combined with a human-centered focus. This doctoral network addresses this critical need by developing a training program to sculpt the next generation of analytics experts combining ML, OR and trustworthy AI and delivers a training framework for others to use and expand. The network includes industrial collaborators from key economic sectors, including logistics, healthcare, public transportation, manufacturing, consulting and communication, to ensure that the methods and knowledge generated in this project are widely applicable and will make an impact both scientifically and in industry.