Simulations of human behavior. The goal is to create efficient, high fidelity simulations of humans in various environments and replicate known behavioral phenomena. This work aims to provide a solid base system for managing NPC actions in video games.
My other goal for this project is to build experience in rust.
- Rust/Cargo
$ cargo run
- world
- real-vector position in finite a 2D plane
- resources
- terrain
- elements/effects
- TBD
- actors
- hunger
- thirst
- energy/fatigue
- injuries
- human armature that simulates reachable workspace, and is effected by injuries
- age/growth
- planning
- Goal Oriented Action Planning (GOAP). What about planning with uncertainty? -> expected outcomes?
- collaboration
- communication
- tools/fabrication
- Information/Knowledge flow
- all information consists of logical propositions. some examples:
- there is a town 1 mile north of here
- character A was alive yesterday
- character B is a friend of character A
- etc.
- all propositions are stored in a central location. Actors can have pointers to particular pieces of knowledge
- actors create new propositions, meaning new knowledge is created. A proposition is delete when no actors have a reference to it.
- actors can transfer knowledge to other actors (depending on things like trust, deception, plausibility, confirmation bias match, etc.)
- GOAP makes use of knowledge. should have probabilistic approach for when knowledge conflicts?
- all information consists of logical propositions. some examples:
- 2D ascii terminal display
- simple text grid
- color