| name | game-ai-optimizer |
| description | Optimizes game AI for strategy games, including troop movement, city defense evaluation, and decision-making. Invoke when user needs to improve AI behavior or fix AI-related issues. |
Game AI Optimizer
This skill provides techniques and best practices for optimizing AI behavior in strategy games, based on practical experience with the Sango Infinity codebase.
Key Optimizations
1. Troop AI Improvements
- Safe Movement: Implemented
MoveToTargetSafely method that evaluates cell safety and terrain bonuses
- Pathfinding: Used
GetDirectPath for reliable pathfinding instead of non-existent methods
- Range Calculation: Ensured
tempMoveRange is properly populated before movement decisions
2. City Defense Evaluation
- Balanced Calculation: Reduced city defense strength weights to prevent部队 from retreating prematurely
- Strategic Decision-making: Removed overly strict defense-based retreat conditions
- Simplified Logic: Streamlined AI decision process to focus on enemy presence rather than static defense values
3. Force AI Enhancements
- Personality System: Added strategy-influencing fields to
Personality class for AI behavior customization
- Technology Research: Fixed AI tech research logic with proper cost calculation and availability checks
- Captive Management: Implemented personality-based tendencies for prisoner handling
4. City AI Optimizations
- Construction Logic: Reduced minimum officer requirement from 3 to 1
- Dynamic Planning: Updated construction logic to check
freePersons.Count instead of assuming fixed values
- Resource Management: Restored construction count limits for better resource allocation
Code Structure Best Practices
- Utility Methods: Centralized AI utility functions in
TroopAIUtility for better code organization
- Type Safety: Fixed type-related issues and ensured proper null checks
- Performance: Optimized pathfinding and movement calculations for better runtime performance
Troubleshooting Tips
- Debugging: Use
GetDiagnostics to identify compilation errors
- Testing: Verify AI behavior by observing troop movements and decision-making
- Balancing: Adjust weight values and thresholds to achieve desired AI behavior
Usage Examples
- When troops are retreating too early: Check defense strength calculations and adjust weights
- When AI gets stuck: Ensure proper pathfinding and movement range calculation
- When AI makes poor decisions: Review personality-based behavior and adjust parameters
