Planning Questions and Answers

Q: What is the overall summary of the planning topic?

Everything in the topic follows from representing states and actions so that the structure is visible, which permits relaxations, heuristics and interaction analysis that atomic states could not support. Exhaustive search is hopeless, planning and learning are complementary, and domain size is not the limit. That a change of representation unlocks the techniques, rather than a cleverer search, is the same lesson constraint satisfaction taught, and it is worth carrying forward.

Q: What is a planning competition?

The competitions run every few years, introduce new domains and compare submitted planners on a common set of problems, which has driven both progress and the standardisation of the description language. Simultaneous execution, industrial evaluation and standardisation are consequences or adjacent activities rather than the definition. A recurring public comparison that sets the field's agenda is a powerful institution, and its choice of domains shapes what gets worked on.

Q: What has the planning description language contributed?

A shared language means a domain written once can be attempted by every planner, and a planner written once can attack every published domain, which is what makes the competitions and the benchmark libraries possible. Guarantees, translation and verification are different matters. An interchange standard whose value lies entirely in what it enables others to do is easy to underestimate and hard to replace once established.

Q: Why do successive versions of the language add features?

Each extension relaxes an assumption of the classical formalism that real domains violate, and planners supporting the richer fragments are correspondingly harder to write. The competitions define tracks by which fragment a planner supports. Inconsistency, easier problems and compatibility are not the drivers. A standard growing in response to the gap between its assumptions and the applications it is asked to serve is the normal life cycle of such a language.

Q: What is a durative action?

A durative action has a start, an end and conditions that may be required at either or throughout, which is what temporal planning adds to the classical formalism. Repetition, persistence and planning cost are different notions. Once actions occupy intervals, two of them may overlap, and the whole question of interference between concurrent actions arises, which is why duration is a substantial extension rather than a numeric annotation.

Q: What is an invariant condition of a durative action?

Flying an aircraft requires fuel throughout rather than only at take off, so the condition must be checked over the whole interval and no concurrent action may violate it. Start and end conditions are the other two kinds. Plan wide invariants and postconditions are different notions. Distinguishing conditions by when they must hold is what allows concurrency to be checked precisely rather than conservatively.

Q: Why must concurrent durative actions be checked for interference?

If one action requires a resource throughout its interval and another consumes it midway, the two cannot overlap even though neither alone is faulty. That check is the temporal analogue of the planning graph's mutex relations. Duration arithmetic, total time and execution agency are different concerns. Interference between overlapping activities is the characteristic difficulty of concurrency wherever it appears.

Q: What is a timeline in a temporal planning system?

Timeline based planners maintain, for each state variable, the intervals over which each value holds, and actions are constraints on those timelines. Spacecraft operations planners are typically built this way. Search order, total duration and resource assignment are different notions. Organising a plan around what is true when, rather than around a sequence of actions, is a different representation that suits domains where many things happen at once.

Q: Why do spacecraft operations use timeline based planning?

Instruments, communications, attitude control and power all evolve simultaneously under strict constraints, so a representation making concurrent evolution explicit is far more natural than a sequence of actions. Communication delay is a genuine reason for autonomy rather than for the representation, and the domains are large. Choosing a representation that matches the structure of the domain, rather than the default of the field, is what makes an application tractable.

Q: Which question does a scheduler answer that a planner does not?

Choosing the actions and their causal order is planning, while assigning start times and resources to already chosen actions is scheduling, and a full application needs both. Agent count, timing of the computation and precision are not the distinction. Separating the causal question from the temporal and resource questions is what allows each to be handed to the technology best suited to it, which is how large applications are actually built.

91.

What is the overall summary of the planning topic?

A.Planning has been superseded entirely by reinforcement learning
B.Planning applies only to domains with fewer than a hundred actions
C.A structured representation enables heuristics that make search feasible
D.Planning problems are solved by exhaustive search of the state space

Answer: C

Everything in the topic follows from representing states and actions so that the structure is visible, which permits relaxations, heuristics and interaction analysis that atomic states could not support. Exhaustive search is hopeless, planning and learning are complementary, and domain size is not the limit. That a change of representation unlocks the techniques, rather than a cleverer search, is the same lesson constraint satisfaction taught, and it is worth carrying forward.

Discuss in Forum

92.

What is a planning competition?

A.A contest between planners running on the same hardware simultaneously
B.An evaluation of planners against real industrial applications
C.A process for standardising the planning description language
D.A periodic event comparing planners on shared benchmark domains

Answer: D

The competitions run every few years, introduce new domains and compare submitted planners on a common set of problems, which has driven both progress and the standardisation of the description language. Simultaneous execution, industrial evaluation and standardisation are consequences or adjacent activities rather than the definition. A recurring public comparison that sets the field's agenda is a powerful institution, and its choice of domains shapes what gets worked on.

Discuss in Forum

93.

What has the planning description language contributed?

A.A common format allowing planners and domains to be exchanged
B.A guarantee that any described problem can be solved efficiently
C.An automatic translation from natural language into planning terms
D.A method for verifying that a domain description is correct

Answer: A

A shared language means a domain written once can be attempted by every planner, and a planner written once can attack every published domain, which is what makes the competitions and the benchmark libraries possible. Guarantees, translation and verification are different matters. An interchange standard whose value lies entirely in what it enables others to do is easy to underestimate and hard to replace once established.

Discuss in Forum

94.

Why do successive versions of the language add features?

A.The language must change to remain compatible with new planners
B.Real applications need durations, resources and conditional effects
C.Earlier versions were found to be logically inconsistent
D.Competitors requested features to make the problems easier

Answer: B

Each extension relaxes an assumption of the classical formalism that real domains violate, and planners supporting the richer fragments are correspondingly harder to write. The competitions define tracks by which fragment a planner supports. Inconsistency, easier problems and compatibility are not the drivers. A standard growing in response to the gap between its assumptions and the applications it is asked to serve is the normal life cycle of such a language.

Discuss in Forum

95.

What is a durative action?

A.An action whose effects persist indefinitely once applied
B.An action that takes longer to plan than to execute
C.An action occupying an interval rather than an instant
D.An action that may be repeated for as long as necessary

Answer: C

A durative action has a start, an end and conditions that may be required at either or throughout, which is what temporal planning adds to the classical formalism. Repetition, persistence and planning cost are different notions. Once actions occupy intervals, two of them may overlap, and the whole question of interference between concurrent actions arises, which is why duration is a substantial extension rather than a numeric annotation.

Discuss in Forum

96.

What is an invariant condition of a durative action?

A.A condition that never changes during the whole plan
B.A condition required only at the moment the action begins
C.A condition guaranteed to hold once the action completes
D.A condition that must hold throughout the action's interval

Answer: D

Flying an aircraft requires fuel throughout rather than only at take off, so the condition must be checked over the whole interval and no concurrent action may violate it. Start and end conditions are the other two kinds. Plan wide invariants and postconditions are different notions. Distinguishing conditions by when they must hold is what allows concurrency to be checked precisely rather than conservatively.

Discuss in Forum

97.

Why must concurrent durative actions be checked for interference?

A.One may violate a condition another requires throughout
B.Concurrent actions always take longer than sequential ones
C.The planner cannot compute the total duration otherwise
D.Durative actions cannot be executed by a single agent

Answer: A

If one action requires a resource throughout its interval and another consumes it midway, the two cannot overlap even though neither alone is faulty. That check is the temporal analogue of the planning graph's mutex relations. Duration arithmetic, total time and execution agency are different concerns. Interference between overlapping activities is the characteristic difficulty of concurrency wherever it appears.

Discuss in Forum

98.

What is a timeline in a temporal planning system?

A.A schedule assigning each action to a particular resource
B.A record of what holds over each interval for a state variable
C.The sequence in which the planner considered the actions
D.The total time the plan will take to execute completely

Answer: B

Timeline based planners maintain, for each state variable, the intervals over which each value holds, and actions are constraints on those timelines. Spacecraft operations planners are typically built this way. Search order, total duration and resource assignment are different notions. Organising a plan around what is true when, rather than around a sequence of actions, is a different representation that suits domains where many things happen at once.

Discuss in Forum

99.

Why do spacecraft operations use timeline based planning?

A.Communication delays make any other approach impossible
B.The domains are small enough for exhaustive enumeration
C.Many subsystems act concurrently under tight resource limits
D.Spacecraft actions are all instantaneous and easy to sequence

Answer: C

Instruments, communications, attitude control and power all evolve simultaneously under strict constraints, so a representation making concurrent evolution explicit is far more natural than a sequence of actions. Communication delay is a genuine reason for autonomy rather than for the representation, and the domains are large. Choosing a representation that matches the structure of the domain, rather than the default of the field, is what makes an application tractable.

Discuss in Forum

100.

Which question does a scheduler answer that a planner does not?

A.A plan applies to one agent and a schedule to several
B.A plan is computed offline and a schedule during execution
C.A plan is approximate whereas a schedule is exact
D.A plan says what to do; a schedule says when and with what

Answer: D

Choosing the actions and their causal order is planning, while assigning start times and resources to already chosen actions is scheduling, and a full application needs both. Agent count, timing of the computation and precision are not the distinction. Separating the causal question from the temporal and resource questions is what allows each to be handed to the technology best suited to it, which is how large applications are actually built.

Discuss in Forum

Loading…

Topics

Other Subjects

Tags: Planning Questions and Answers || Planning MCQ Questions and Answers || Planning GK Questions and Answers || Planning GK MCQ Questions || Planning Multiple Choice Questions and Answers || Planning GK || GK on Planning || Artificial Intelligence Questions and Answers || Artificial Intelligence MCQ Questions and Answers || Artificial Intelligence GK Questions and Answers || GK on Artificial Intelligence

Planning Questions and Answers

Topics

Other Subjects

About Us

Stay informed about all important Dates & Notifications