What is the Frame Problem?
how to represent things which stay unchanged after performing some action
-> a large number of so-called frame axioms would be needed in general to represent what does not change by performing an action.
What is the the ramification problem?
the representation of implicit effects
e.g. if a car moves from one position to another, so does – any person in the car, the engine of the car, any dust particle in the car, any bacteria in the driver, etc.
What is the qualification problem?
required preconditions (the “qualifications”) ensuring that an action succeeds
What types are the classical planning environments?
fully observable
deterministic
finite
static (change happens only when the agent acts)
discrete (in time, actions, objects, and effects)
which kind of literals are allowed in STRIPS?
literals can be propositional or first-order, but first-order literals must be ground (i.e., variable-free) and function-free.
How is the precondition formulated?
The precondition:
a conjunction of function-free positive literals stating what must be true in a state before the action can be executed.
How is the effect formulated?
The effect: a conjunction of function-free literals (can be positive and negative) describing how the state changes when the action is executed.
how is the result of executing an applicable action a is a state s′ that results from s displayed?
What are differences between ADL and STRIPS?
What is the forward state-space search?
(or progression planning): from initial state to goal
What is the backward state-space search?
(or regression planning): from goal to initial state
what is the main advantage of backward search?
allows to consider only relevant actions
(An action is relevant to a conjunctive goal if it achieves one of the conjuncts of the goal)
What are examples of totally ordered plan searches? What does it mean?
Forward and backward state-space search are particular forms of totally ordered plan searches.
-> They explore only strictly linear sequences of actions and do not take advantage of problem decomposition.
What is the difference between partial-order and total-order plans?
How does the pop algorithm work?
How does the partial-order-plan algorithm enforce consistency? How does it resolve conflicts?
what does the pop goal test check?
The goal test checks whether a plan is a solution to the original planning problem.
-> Because only consistent plans are generated, the goal test just needs to check that there are no open preconditions.
Last changed2 years ago