AI Vocabulary Exam 1
What does a search problem consist of?
- A state space S - An initial state so - Actions A(s) - Transition model Successor (s,a) - Goal Test G(s) - Action cost c(s,a,s')
What is incorporated in a search tree?
- Nodes represent plans for reaching states - Plans have costs (sum action costs)
What are the different types of environments?
- Partially observable - stochastic - Multi-agent - static - continuous time - unknown physics - unknown performance measure
What are the different agent types? In order of increasing generality and complexity
- Simple reflex agents - Reflex with state - Goal-based agents - Utility-based agent
What is a search problem
- States (configurations of the worlds) - Actions and costs - Successor function (world dynamics) - Start state and goal test
What does a search algorithm have?
- Systemically builds a search tree - Chooses an ordering of the frontier (unexplored nodes) - Optimal: finds least-cost plans
What is a heuristic h is admissible (optimistic)?
A heuristic is admissible (optimistic) if: 0 <= h(n)<=h*(n) where h*(n) is the true cost to a nearest goal
What is a planning agent?
A planning agent is an agent that ask questions and formulates decisions based on (hypothesized) consequences of actions. Must have a model of how the world evolves in response to actions and it must formulate a goal.
What is a rational agent?
A rational agent chooses actions that maximizes the expected value of the performance measure
What is a reflex agent?
A reflex agent is an agent that chooses actions based on current percept (and maybe memory) they might contain memory or a model of the world's current state but they do not consider the future consequences of their actions.
What is a search state?
A search state keeps only the detail needed for planning (abstraction)
What is a solution?
A solution is a sequence of actions that reaches a goal state
What is the actuators in PEAS?
Actuators is the part of the agent that delivers the output of an action to the environment
What is the agent design in a static environment?
Agent has time to compute a rational decision
What is the agent design in the stochastic environment?
Agent may have to prepare for contingencies
What is the agent design do in a multi-agent environment?
Agent may need to behave randomly
What is the agent design in a partially observable environment?
Agent requires memory (internal state)
What is an admissible heuristic?
An admissible heuristic is used to estimate the cost of reaching the goal state in an informed search algorithm
How does an agent perceive its environment?
An agent perceives its environment through its sensors and acts upon it through its actuators or effectors
What is an agent?
An entity that perceives or acts
What is an optimal solution
An optimal solution has the least cost among all solutions
What are the spectrum of representation?
Atomic, factored, structured
What is the agent design in a continuous environment?
Continuously operating controller
In a state's space graph how many times does each state occur?
Each state occurs only once
What is environment in PEAS?
Environment is the surroundings of the agent at every instant. It keeps changing with time if the agent is set in motion.
What is the agent design in an unknown physics environment?
Need for exploration
Can every agent function be implemented by some agent program?
No
Do rational agents make mistakes?
No, but they can be unsuccessful
Are rational agents omniscient(all knowing)?
No, they are limited by the available percepts
Are rational agents clairvoyant (ability to perceive events in the future)?
No, they lack the knowledge of environment dynamics
What is PEAS function?
PEAS description defines the task environments; precise PEAS specifications are essential and strongly influence agent design
What is the performance measure in PEAS?
Performance measure is the unit to define the success of the agent
What does PEAS stand for?
Performance measure, Environment, Actuators, Sensors
Replanning vs. Preplanning
Replanning is when they make a decision and keep making decisions as they go. Agents know what to do when something unexpected happens. Preplanning is when the plan their course of actions before they move
What is sensors in PEAS?
Sensors are the receptive part of the agent which takes in the input for the agent
What does the agent function do?
The agent functions maps from percept histories to actions f: P* ->A
What does the agent program i run on?
The agent program i runs on some machine M to implement f: f = Agent(I, M)
What is a world state?
The world state includes every last detail of the environment
Are rational agents autonomous?
Yes, as they learn, their behavior depends more on their own experience
Do rational agents explore and learn?
Yes, in unknown environments these are essential
Can a reflex agent be rational?
Yes, it can be
What is the agent design in an unknown performance measure environment?
observe/interact with human principal
