Planning as a CSP

In forward planning, the nodes considered are constrained

to be reachable, even if they don’t lead to goal.

Artificial Intelligence, Lecture 6.4 1/

Planning as a CSP

In forward planning, the nodes considered are constrained to

be reachable, even if they don’t lead to goal.
In regression planning, the nodes considered are constrained
to be ones from which we can achieve the goal, even if they
are not reachable.

Artificial Intelligence, Lecture 6.4 1/

Planning as a CSP

In forward planning, the nodes considered are constrained to

be reachable, even if they don’t lead to goal.
In regression planning, the nodes considered are constrained
to be ones from which we can achieve the goal, even if they
are not reachable.
When representing planning as a CSP, we can constrain the
states by the starting state and the goal.

Artificial Intelligence, Lecture 6.4 1/

Planning as a CSP

In forward planning, the nodes considered are constrained to

be reachable, even if they don’t lead to goal.
In regression planning, the nodes considered are constrained
to be ones from which we can achieve the goal, even if they
are not reachable.
When representing planning as a CSP, we can constrain the
states by the starting state and the goal.
. . . but we can only do this if we know the number of
steps.

Artificial Intelligence, Lecture 6.4 1/

Planning as a CSP

In forward planning, the nodes considered are constrained to

be reachable, even if they don’t lead to goal.
In regression planning, the nodes considered are constrained
to be ones from which we can achieve the goal, even if they
are not reachable.
When representing planning as a CSP, we can constrain the
states by the starting state and the goal.
. . . but we can only do this if we know the number of
steps.
Search over planning horizons (number of time steps).

Artificial Intelligence, Lecture 6.4 1/

Planning as a CSP

In forward planning, the nodes considered are constrained to

be reachable, even if they don’t lead to goal.
In regression planning, the nodes considered are constrained
to be ones from which we can achieve the goal, even if they
are not reachable.
When representing planning as a CSP, we can constrain the
states by the starting state and the goal.
. . . but we can only do this if we know the number of
steps.
Search over planning horizons (number of time steps).
For each planning horizon, create a CSP constraining
possible actions and features

Artificial Intelligence, Lecture 6.4 1/

CSP Variables

Choose a planning horizon k.

Create a variable for each state feature and each time from
0 to k.
Create a variable for the action for each time in the range 0
to
k − 1.

Artificial Intelligence, Lecture 6.4 2/ 6

CSP for Delivery Robot for a planning horizon of
2
RLoc0 RLoc1 RLoc2

RHC0 RHC1 RHC2

Action0 Action1

SWC0 SWC1 SWC2

MW0 MW1 MW2

RHM0 RHM1 RHM2

State1
State0 Action0 Action1

RLoci — Rob’s location State2

RHCi — Rob has coffee

SWCi — Sam wants coffee Movei — Rob’s move action
MWi — Mail is waiting PUCi — Rob picks up
coffee DelC — Rob delivers
RHMi — Rob has mail
coffee PUMi — Rob picks
Artificial Intelligence, Lecture 6.4 3/ 6
Constraints

precondition constraints between state variables at time t

and action variable at time t, specify constraints on what
actions are available from a state.

Artificial Intelligence, Lecture 6.4 4/

Constraints

precondition constraints between state variables at time t and

action variable at time t, specify constraints on what actions
are available from a state.
effect constraints between state variables at time t, action
variable at time t and state variables at time t + 1
constrain the resulting state to be one that satisfies the
effects.

Artificial Intelligence, Lecture 6.4 4/

Constraints

precondition constraints between state variables at time t

and action variable at time t, specify constraints on what
actions are available from a state.
effect constraints between state variables at time t, action
variable at time t and state variables at time t + 1
constrain the resulting state to be one that satisfies the
effects.
frame constraints among state variables at time t, action
variables at time t, and state variables at time t + 1 for
values of variables that do not change.

Artificial Intelligence, Lecture 6.4 4/

Constraints

precondition constraints between state variables at time t

and action variable at time t, specify constraints on what
actions are available from a state.
effect constraints between state variables at time t, action
variable at time t and state variables at time t + 1
constrain the resulting state to be one that satisfies the
effects.
frame constraints among state variables at time t, action
variables at time t, and state variables at time t + 1 for
values of variables that do not change.
initial state constraints that are usually domain constraints
on the initial state (at time 0).
goal constraints that constrains the final state to be a
state that satisfies the goals that are to be achieved.

Artificial Intelligence, Lecture 6.4 4/

Example Constraints

precondition constraint RLoci = cs if Actioni = puc

is violated when

Artificial Intelligence, Lecture 6.4 5/

Example Constraints

precondition constraint RLoci = cs if Actioni = puc

is violated when RLoci /= cs ∧ Actioni =
puc

Artificial Intelligence, Lecture 6.4 5/

Example Constraints

precondition constraint RLoci = cs if Actioni = puc

is violated when RLoci /= cs ∧ Actioni =
puc
effect constraint rhci +1 if Acti = puc
is violated when

Artificial Intelligence, Lecture 6.4 5/

Example Constraints

precondition constraint RLoci = cs if Actioni = puc

is violated when RLoci /= cs ∧ Actioni = puc
effect constraint rhci +1 if Acti = puc
is violated when RHCi +1 = false ∧ Acti =
puc.

Artificial Intelligence, Lecture 6.4 5/

Example Constraints

precondition constraint RLoci = cs if Actioni = puc

is violated when RLoci =/ cs ∧
Actioni = puc
effect constraint rhci +1 if Acti = puc
is violated when RHCi +1 = false ∧ Acti = puc.
frame constraint Rob has mail at any time if it had mail
before and the action wasn’t

Artificial Intelligence, Lecture 6.4 5/

Example Constraints

precondition constraint RLoci = cs if Actioni = puc

is violated when RLoci /= cs ∧ Actioni = puc
effect constraint rhci +1 if Acti = puc
is violated when RHCi +1 = false ∧ Acti = puc.
frame constraint Rob has mail at any time if it had mail before
and the action wasn’t to pickup mail or deliver
mail:

RHMi +1 = RHMi if Acti ∈/ { pum, dm}

violated when

Artificial Intelligence, Lecture 6.4 5/

Example Constraints

precondition constraint RLoci = cs if Actioni = puc

is violated when RLoci /= cs ∧ Actioni = puc
effect constraint rhci +1 if Acti = puc
is violated when RHCi +1 = false ∧ Acti = puc.
frame constraint Rob has mail at any time if it had mail before
and the action wasn’t to pickup mail or deliver
mail:

RHMi +1 = RHMi if Acti ∈/ { pum, dm}

violated when
RHMi +1 /= RHMi ∧ Acti /= pum ∧ Acti /= dm

Artificial Intelligence, Lecture 6.4 5/