Moving AI Lab

Nathan R. Sturtevant
Canada CIFAR chair
Alberta Machine Intelligence Institute (Amii) Fellow
Professor, Department of Computing Science, University of Alberta

---

The Moving AI lab does research broadly in Artificial Intelligence. This includes work that includes but is not limited to heuristic search (single-agent, multi-agent, game-tree), learning (in search and traditional games), game playing, and video games. Our pathfinding research has been implemented in commercial games (Dragon Age: Origins and the upcoming Nightingale).

---

Broad resources of community interest:

• Benchmark problems for:
  • 2D grids
  • Voxel grids
  • Multi-agent Path Planning (MAPF)
• Historical Grid-Based Path Planning Competition data. (We are working to revive this, but plans were disrupted by COVID-19.)
• Interactive heuristic search demos. These cover many basic and advanced heuristic search algorithms. These demos and other research is implemented in the HOG2 platform.
• Graduate lectures on Single Agent Search

---

Recent additions of interest:

• Exhaustive Procedural Content Generation: Snakebird
• Heuristic Search Demos:
  • Budgeted Tree Search: IDA* replacement with better worst-case performance
  • Weighted A* (XDP): Simple modification to Weighted A* that often finds suboptimal solutions more efficiently, especially on hard problems