Environment

In General

Each architecture was analysed in terms of the environments it could live in. The environment can be viewed from several dimensions: Amount of dynamic behavior, complexity, predictability of changes to environment, whether the environment is real or simulated, tolerance to noise in sensors and actuators, and tolerance of imperfect world-state knowledge.

*** put unnumbered lists back in for Dynamic/Not, etc.

Dynamic/static

Can the environment change independent of the situated agent?

Dynamic

Not dynamic (or severly limited dynamicity)

Complexity

Can the agent function in complex environments?

Complex

Simple (or limited complexity)

Predictablility

Must changes in the environment be predictable by the agent?

Predictablity required

Unpredictablity allowed

Real/simulated

In what types of environments has the agent been tested or developed?

Real

Simulated

Noise Tolerance

In what tolerance does the agent have to noise in the environment or sensors?

Noise Tolerant

Little or no noise tolerance

Noise not an issue

Perfect Knowledge Requirements

Can the agent deal with imperfect knowledge gracefully?

Tolerant of imperfect knowledge

Knowledge not an issue

Perfect knowledge Required