Building the AI for Athena Crisis - Christoph Nakazawa, JS GameDev Summit 2023

Ai

Discover the simplicity and power of building AI for Athena Crisis, a stateless and composable system designed for ease of implementation and maintenance, with a search algorithm and basic action function for optimal gameplay.

Key takeaways
  • The AI system for Athena Crisis is designed to be simple, fast, stateless, deterministic, and composable, making it easy to implement and maintain.
  • The AI uses a basic action function to determine the available actions for a unit, such as moving or attacking.
  • The system uses a search algorithm to determine the best action to take, such as finding the closest target or going through the river.
  • The AI doesn’t have memory or planning, it just makes its decision based on the current game state.
  • The system uses a map state to keep track of the game state and units’ positions.
  • The AI can add randomness to make the gameplay more interesting.
  • The system uses an NPM package called sk-means for clustering and grouping units.
  • The AI can attack other units, capture buildings, and move units based on the game state.
  • The system uses a simple architecture with a map state, units, and actions, making it easy to implement and maintain.
  • The AI can be made more aggressive or defensive by adjusting its priorities.
  • The system can use a Fog of War system to keep track of what units can see and what’s hidden.
  • The AI can move units between different positions based on the game state.
  • The system can use a unit data structure to keep track of unit positions and abilities.
  • The AI can attack units with different types of attacks, such as regular attacks or special abilities.
  • The system can use a game state to keep track of the current game situation and units’ positions.