Graphs within graphs - a story of developing spatial analysis tools for the needs of game AI
When developing AI for a turn-based tactical game with continuous space, we encountered many problems involving spatial reasoning and navigation.
This talk will describe the process of detecting and solving those issues with techniques such as annotated graphs, hierarchical pathfinding and spatial regionalization.
Custom-built tools such as perception system and mass maps are also discussed.
Daniel graduated from Computer Science at the Warsaw University of Technology. He started his career by implementing a PC version of tabletop war games. Since 2017, associated with the QED Software group, where he is responsible for R&D related to artificial intelligence in games (including the complex AI middleware – Grail) and the design and implementation of AI algorithms in commercial projects, both in-house and for third-parties.
Since 2014 he’s been an active member of the KNTG Polygon.