Planning (Powerpoint) Flashcards

1
Q

Vilka 4 kategorier av problem finns det?

A
  1. Deterministic, fully observable
    • Single state problems. The agent is aware of its exact state and the solution is a sequence of actions
  2. Non-observable
    • Sensorless problems, solution is a sequence of problems
  3. Non-deterministic
    - percepts offer new insights into the current state, often requiring a combination of search and execution actions
  4. Unknown state space
    - Exploration problems, these involve navigating through a state space where the full space is not initially known
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Vad är en deterministic algorithm?

A

För en given input, kommer datorn alltid att producera samma output som går igenom samma states

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Vad är en non-deterministic algorithm?

A

För samma input, kan kompileraren producera olika output i olika runs. Kan ge olika output från samma input.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Vad är NP-fullständig problem?

A

Non-deterministic Polynomial problem.

  • Klass av matematiska problem för vilka effektiva lösningar saknas
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Vad är den enda lösningen man funnit på ett godtyckligt NP-fullständigt problem?

A

Att i princip gå igenom alla tänkbara lösningar och jämföra dem, vilket är ogenomförbart för andra än små probleminstatser.

Tiden för att verifiera en lösning är storleken på indata.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Vad är constraint Programming?

A

En paradigm i artificiell intelligence för att lösa combinatorial problems.

  • A prblem is modeled as a set of constraint that specify relations between variables. The goal is to find a solution that satisfies all the constraints.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Representations in planning

A

Planering använder logik för att representera
- Actions
- States
- Goals

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Vad är ett språk för att representera planering

A

STRIPS
Används för att representera och lösa planning problems i AI

Representerar state
- Decompose the world in logical conditions and represent a state as a conjunction of positive literals

Representatin of Goals
- Partially specified state and represented as a conjunction of positive ground literals

Representation of Actions
- Action = Precondition + Effect

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Vad är två challenges of AI and Plannning

A
  1. closed world assumption:
    • Assumes that world model contains everything the robot needs to know there can be no surprise
  2. Frame Problem:
    How to represent real world situations in a manner that is computationally tractable
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Hur kan man planera med state-space search?

A

Både forward och backward search är möjlig

  1. Progression planners
    • Forward state-space search
    • Consider the effect of all possible action in a given state
  2. Regression Planners
    - Backward state-space search
    • To achieve a goal, solve what must have been true in the previous state

Progression and regression planning are totally ordered plan search forms

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Vad är partial-order Planning

A

Any planning algorithm that can place two actions into a plan without which comes first

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q
A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly