AASS Seminar: Professor Roman Barták, Charles University, Prague

Who: Professor Roman Barták, Charles University, Prague
Title: Some Thoughts on Robustness in Multi-Agent Path Finding
When: Friday December 12, 10:15
Where: T135, Teknikhuset

Abstract: Multi-agent Path Finding (MAPF) is a problem of finding collision-free paths for a set of agents (robots) moving in a shared environment. It has numerous practical applications, with warehouse robots being the most prominent one. In this talk, we introduce the MAPF problem and discuss how to handle environmental uncertainty during both execution and planning. In particular, we will introduce methods for robust execution of plans and for generating robust plans, such as k-robustness, contingency plans, and using policies instead of plans.

Bio: Roman Barták is a professor of computer science at Charles University, Prague. He works in the area of Artificial Intelligence with particular emphasis on automated planning and scheduling, constraint satisfaction, and knowledge representation. His research goal is to design intelligent autonomous agents (e.g., robots) that can plan their activities, make informed decisions, and act effectively in real-world environments. He put strong emphasis on model-based approaches to problem-solving as a means to achieve trustworthy and explainable AI. Recently, his research has focused on two areas: multi-agent path finding and hierarchical planning.