agent types (slides2)

Question 1

four types of agents

Answer 1

simple reflex, model based reflex, goal based, utility based

Question 2

simple reflex agent code

Answer 2

function SIMPLE-REFLEX-AGENT(percept) returns an action

persistent: rules, a set of condition-action pairs

state <– rule.ACTION return action

action only depends on percept

Question 3

model based reflex agent

Answer 3

• to cope with partial observability the agent keeps track of the part of the world it cannot see
• Internal state of environment depends on sequence of percepts
• model: how the world evolves independently of the agent, how the actions of agent affect the world
• current state usually not exact
• described succinctly

Question 4

model-based reflex agents code

Answer 4

function MODEL-BASED-REFLEX-AGENT(percept)
returns an action
persistent:
state, the agent’s current conception of the world state
model, description of how next state depends on current state and action
rules, set of condition-action pairs
action, the most recent action, initially none

state <– rule.ACTION
return action

Question 5

search and planning

Answer 5

important for model-based and Goal-based agents,
seek action sequences that achieve goals.
flexible, since they dont require rewriting condition rules when the goals change

Question 6

utility-based agents

Answer 6

not all ways to achieve goals are equally good
Goals enable distinguishing between satisfied an unsatisfied
UTILITY FUNCTION represents performance measure of level of satisfaction
Utilities are when there are conflicting goals, or many possible goals
Rationality maximizes expected utility

Question 7

Learning Agents

Answer 7

Better than programming intelligence directly

ability to cope directly with unknown environments

Question 8

Learning Agent Components

Answer 8

learning element
performance element
critic
possible utility
problem generator

Question 9

learning element

Answer 9

responsible for making improvements

-can change any component of the agent

Question 10

performance element

Answer 10

the actual agent

Question 11

critic

Answer 11

assesses how well the agent is performing w.r.t. a fixed performance standard (e.g. checkmate)

Question 12

problem generator

Answer 12

generates actions that yield an information gain(otherwise the ones that are best given the current knowledge would always be chosen)
-May suggest suboptimal actions which lead to find better actions in the long run

Question 13

Utility function

Answer 13

Represents performance measure of level of satisfaction