Abstract
Abstract
MAPF is a fundamental challenge that is frequently encountered in robotics, warehouse automation, and systems with more than one robot. The goal is to find paths for multiple agents that don't collide with each other while also making the best use of global performance criteria. Even though optimal MAPF algorithms have strong theoretical guarantees, they can't be used in situations where time is important and the cost of running them is high.
In real-world situations, it is often more important to find solutions that are possible and of high quality within a set amount of time than to find the best solution. Prioritized planning approaches meet this need by planning agents one at a time based on a set priority order. However, their performance depends heavily on how priorities are chosen and conflicts are resolved.
This thesis proposes a temporal slack-aware prioritized planning framework for MAPF under strict time constraints. Temporal slack is a simple heuristic that shows how much an agent can handle delays and waiting actions without making the solution much worse. The proposed method makes priority assignment and conflict handling more robust and stable in dense and constrained environments by adding temporal slack to them.