VQ is one of the few games to feature solid modeling. It is currently 100 percent procedurally-generated and by default has 2 million voxels per cubic meter.
A fully abstracted, procedurally generated world that is designed to be easy for AI to comprehend.
With the exception of a few effects, every single entity in the game is rendered with voxels and is completely solid.
AI abilities include pathfinding, reconnaissance/knowledge gathering, deduction (propositional logic, queries, backwards chaining), planning, collaboration, and score maximization (where the highest score is represented by the most motivators with the fewest deterrents). The AI does not perform abstraction/recognition as the world is pre-generated and pre-abstracted with this in mind. As a rough example, the AI does not need to look at a group of trees and process it to determine it is a forest. The world is procedurally generated from the top down and the AI understands all higher-order relationships. So, the AI can easily understand that a room is part of a given building which is part of a given city, which is part of a given province, which is part of a given country. In other words, the AI has a built in understanding of the world on macroscopic and microscopic levels. At a high level, the AI functions in many ways like a strategy game / RTS AI. AI knowledge is often approximated as doing a realtime world-wide simulation is obviously way too computationally expensive. The AI can gather existing knowledge, but it does not use a learning system (i.e. a weighted neural network or other similar mechanism).
This page was last edited on 30 October 2014, at 21:09.
Content is available under CC BY-NC-SA 3.0 unless otherwise noted. Game content and materials are trademarks and copyrights of their respective publisher and its licensors. All rights reserved.
This site is a part of Fandom, Inc. and is not affiliated with the game publisher.