Needs-based AI is a general term for action selection based on attempting to fulfill a set of mutually competing needs. An artificial agent is modeled as having a set of conflicting motivations, such as the need to eat or sleep, and the world is modeled as full of objects that can satisfy those needs at some cost.
The core of the AI is an action selection algorithm that weighs those various possible actions against each other, trying to find the best one given the agent’s needs at the given moment. The result is a very flexible and intuitive method for building moderately complex autonomous agents, which are nevertheless efficient and easy to understand.
This gem presents some technical details of needs-based ...