Highlights
In this episode of Postbot, Drekken and co-host Ketroc chat with Simon and Geertien about their bot Tyr and delve into the intricacies of heat maps and influence maps. They discuss the challenges and strategies of implementing these maps to improve bot behavior, particularly in pathfinding and combat scenarios.
Reusability of Code: Simon and Geertien emphasize the importance of reusing code for functionalities such as pathfinding to avoid enemy units. They discuss how influence maps can help bots make better decisions by understanding the danger levels of different map regions.
Framework Comparison: The conversation touches on the advantages of using influence maps over traditional methods, highlighting how pre-calculated paths and danger zones can enhance bot performance in competitive scenarios.
Focus on Efficiency and Optimization: They discuss the need for efficient pathfinding algorithms, like Dijkstra’s algorithm, and the potential for optimizing these processes by calculating partial paths and grouping units.
Advanced Features Support: Simon and Geertien talk about the potential of influence maps in scenarios like drops and combat, where bots can use these maps to make more strategic decisions such as avoiding heavily guarded areas or maximizing damage output.
Strategic Development Encouragement: The discussion includes how influence maps can encourage more nuanced bot behavior by abstracting complex decision-making processes, allowing developers to focus on higher-level strategies and tactics.