Agents and Environment - Concept of Rationality - Nature of Environment

Principles of Artificial Intelligence 21AI54
MODULE – 1
1. Agents and Environment
➢ Agent: is anything that can be viewed as perceiving its environment through

Sensors & acting upon that environment through Actuators.
Humans: Eyes, Ears (Sensors)

Hand, Leg (Actuators)
Robotic Agent: Cameras, Infrared finders (Sensors)

Motors (Actuators)
➢ Mathematically, an Agent’s behaviour is described by Agent Function, that maps any

given percept sequence to an action.
➢ Internally, Agent function for an Artificial agent will be implemented by an Agent
program.
Example: A Vacuum Cleaner world
- Can invent many variations square

- This world has two locations A and B
- Vacuum agent perceives which square it is in and whether there is dirt in the
square.
- It can choose to move Left, move Right, Suck dirt, or do Nothing.
- Agent Function: if current square is dirty, then suck. Otherwise move to another
square.
Function vacuum_agent
If status = Dirty, then return Suck
Else if Location = A, then return Right
Else if Location = B, then return Left
Partial Tabulation of Simple Agent Function
2. The Concept of Rationality
➢ Rational Agent: is one that does the right thing

➢ If the sequence is desirable, then the agent has performed well.
➢ Desirability is captured by Performance Measure that evaluates any given sequence
of environment states.
➢ As a general rule, it is better to design Performance Measures according to what one
actually wants in the environment rather than according to how one think the agent
should behave.
[2.1] Rationality
➢ Rationality at any given time depends on 4 things;
a) Performance Measures that defines the Criterion of Success.
b) Agent’s prior knowledge of the environment.
c) Actions that the agent can perform.
d) Agent’s percept sequence to date.
➢ Definition of Rational Agent: For each possible percept sequence, a rational
agent should select an action that is expected to maximize its performance
measure, given the evidence provided by the percept sequence & whatever
built-in knowledge the agent has.
[2.2] Omniscience, learning & autonomy

➢ An omniscient agent knows the actual outcome of its actions & can act
accordingly but omniscience is impossible in reality.
➢ Rationality maximizes expected performance while Perfection maximizes
actual performance.
➢ Doing actions in order to modify future percepts are called Information

Gathering.
➢ A rational agent should be Autonomous i.e it should learn what it can to
compensate for partial or incorrect prior knowledge.
3. The Nature of Environments

➢ To build a Rational Agent, we need to think about the Task Environment, which
are essentially the problems & the rational agents are the solutions.
[3.1] Specifying the task environment

✓ While understanding Rationality, we had come across the following;
a) Performance
b) Environment
c) Actuators
d) Sensors
In short we call it as PEAS Description.
✓ In designing an agent, the first step is to specify the task environment.
Example: PEAS description for an automated taxi
1) What is the performance measure to which we would like our automated driver
to aspire?
- Getting to correct destination

- Minimize fuel consumption
- Minimize traffic violations
- Maximizing passenger safety & comfort
2) What is driving environment?

- Deal with variety of roads
- Pedestrians, stray animals, police cars
- potholes
3) Actuators for automated taxi
4) Sensors for automated taxi
[3.2] Properties of Task Environments
a) Fully observable / Partially observable

- Fully: if agent’s sensor gives full access to complete state of the environment.
- Partial: environment might be partially observable because of noisy & inaccurate
sensors.
b) Single agent/ Multi-agent
Deterministic: Solving a crossword puzzle
Multi-agent: Playing chess game
c) Deterministic/ Stochastic
If an agent is able to determine the next state from the current state, & action
executed by agent.
d) Episodic/ Sequential
Agent’s experience is divided into atomic episodes. In each episode, the agent
receives a percept & then performs a single action.
e) Static/Dynamic
if environment can change while an agent is deliberating, then its Dynamic.
f) Discrete/ Continuous
This distinction applies to the state of the environment to the way time is handled
and to the percepts & action of the agent.
g) Known/Unknown
This distinction refers not only to the environment, but to the agent’s state of
knowledge about the “laws of physics” of the environment.
In known environment, the outcomes for all actions are given.
In unknown environment, the agent will have to learn how it works in order to
make good decisions.

Agents and Environment - Concept of Rationality - Nature of Environment

Uploaded by

Copyright:

Available Formats

Agents and Environment - Concept of Rationality - Nature of Environment

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Agents and Environment - Concept of Rationality - Nature of Environment

Uploaded by

Copyright:

Available Formats

Principles of Artificial Intelligence 21AI54

1. Agents and Environment

➢ Agent: is anything that can be viewed as perceiving its environment through

Humans: Eyes, Ears (Sensors)

Robotic Agent: Cameras, Infrared finders (Sensors)

➢ Mathematically, an Agent’s behaviour is described by Agent Function, that maps any

Example: A Vacuum Cleaner world

- Can invent many variations square

2. The Concept of Rationality

➢ Rational Agent: is one that does the right thing

[2.2] Omniscience, learning & autonomy

➢ Doing actions in order to modify future percepts are called Information

3. The Nature of Environments

[3.1] Specifying the task environment

Example: PEAS description for an automated taxi

- Getting to correct destination

2) What is driving environment?

[3.2] Properties of Task Environments

a) Fully observable / Partially observable

You might also like