Battle Module
· 2 min read
I came up with a battle module for Norchevsky
Key Features?
- Positioning in space and time of all objects (characters, items, covers, etc.)
- Tracking and aggregation of all objects' state
- Convenient access to tracking data
- Conducting the battle itself
Battle Management Ideas
- Round Dance
- Swarm
- Hybrid
There is one main agent that programmatically calculates all possible decisions and, based on this data and orders from above, determines the behavior of battle participants.
Pros
Good dynamics due to comprehensive non-player combat design
Cons
- Hyper-factorial complexity
- Too many calculations required
- Risk of the round dance agent being overwhelmed by data
Each battle participant has their own agent (and AI), their own memory, and is fully autonomous. The battle is conducted in turn-based mode.
Pros
- Easier to implement due to turn-based system and local processing
- Can create more scenarios (e.g., when an agent fails to notice something)
Cons
Slightly reduced dynamics, but since AI processes quickly, turns will be returned to the player promptly.
A blend of both previous approaches, attempting to combine their strengths.
Differs from the swarm by having a central coordinator that reminds agents of common goals, tactics, and other pre-agreed battle parameters.