Parallel AI Planning

AI planning is a core challenge in artificial intelligence. It focuses on intelligent sequential decision-making to achieve specific objectives within a given environment.

State-of-the-art AI planning algorithms can solve complex planning tasks from various application domains, such as robotics, logistics, and space mission control.

However, while adjacent AI fields have successfully embraced parallel computing, AI planning algorithms are still constrained to purely sequential, single-core execution.To address this, I will create the first parallel AI planning system, fully optimized for modern multi-core processors. Within it, I will develop new parallel search algorithms tailored for AI planning tasks.

These algorithms will greatly speed up the search for plans, but the speedup will be limited by the number of available cores.

To achieve an exponential reduction in search effort, the main goal of my project is to develop the first algorithms for computing search guidance in parallel.

By combining solid theoretical foundations with efficient parallel implementations, my project has the potential to induce a paradigm shift in AI planning research towards parallel computing.