Real-time OR, Baseline Knowledge, and More: Our Conversation with Filippo Focacci

In our latest conversation on the Decision Intelligence Lab Podcast, Vijay and I hosted Filippo Focacci. I’ve been friends with Filippo since we worked together at ILOG and IBM. He has a lot to say on solving hard optimization problems.

As an aside, he mentioned that his retirement career may be to open a restaurant. My family had the chance to visit him in his hometown in southern France (picture below), and I can attest that he is both a great cook and a great host.

But, back to the podcast…

Here were the highlights for me:

Interesting real-time applications. He talked about using optimization for a bike-sharing organization (rebalancing bikes) and real-time scheduling of airport cleaning crews (for example, if a flight is late, there is no way to clean the gate with people there, and once the flight eventually leaves, cleaning will take longer).

Improving planning at Toyota with good baseline models. Toyota is good at planning. Although the process Filippo worked on was manual, it contained a lot of knowledge. Filippo discusses the importance of first replicating the current process and then improving it with optimization.

ROI comes with execution. He made a great point that you get savings in the execution of the plan, not in the plan.

Building a team. He offers good advice for building an optimization team and discusses various optimization technologies required to address the problem.

Yves Caseau (Group Chief Digital & Information Officer at Michelin), GenAI, and a system of systems. We had a good discussion on some of the ideas of Yves and how to think about GenAI.

I hope you enjoy the episode.