Keeps track (i.e. a model) of the part of the world it can’t see. Requires some internal state that depends on percept history. Transition model: Knowledge of how the world changes based on the agent actions or external events. Sensor model: knowledge about how the sensors reflect the world.
The model of the world is not usually perfect, just a best guess.
function MODEL-BASED-REFLEX-AGENT(percept) returns an action
persistent: state, the agent's current conception of the world state
model, a description of how the next state depends on current state and action
rules, a set of condition-action rules
action, the most recent action, initially none
state ← UPDATE-STATE(state, action, percept, model)
rule ← RULE-MATCH(state,rules)
action ← rule.ACTION
return action
This is still a reflex-based agent but instead of just the current percept, it operates based on its knowledge of the world.
This fixes the issue of needing full observability, but we still need a set of condition-actions (if-else statements) to determine what to do.
Learning Agents