Scribd
Upload
1K views30 pages

Partial Order Plan

Partial order planning works by developing subplans independently and combining them, allowing for flexibility in action ordering. It considers putting on socks and shoes as independent subplans that can be combined. Planning graphs are used for heuristic estimation in partial order planning, representing states and possible actions at each time step to estimate goal achievement. The GraphPlan algorithm alternates between extracting solutions by searching backwards and expanding the graph to add new actions and states.

Uploaded by

sayedu
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
1K views30 pages

Partial Order Plan

Partial order planning works by developing subplans independently and combining them, allowing for flexibility in action ordering. It considers putting on socks and shoes as independent subplans that can be combined. Planning graphs are used for heuristic estimation in partial order planning, representing states and possible actions at each time step to estimate goal achievement. The GraphPlan algorithm alternates between extracting solutions by searching backwards and expanding the graph to add new actions and states.

Uploaded by

sayedu
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 30

Planning

• Planning techniques have been applied in a


variety of tasks including robotics, process
planning, web-based information gathering,
autonomous agents and spacecraft mission
control.
• Forward and backward state-space search are
forms of totally ordered plan search.
• They explore only strictly linear sequences of
actions directly connected to the start or goal
state.
• They cannot take advantage of problem
decomposition.
Partial Order Planning
• Partial order planning works on several subgoals
independently, solves them with several subplans,
and then combines the subplans.
• It has the advantage of flexibility in the order in which
the planner constructs the plan
• Chronological ordering is not necessary
• The strategy of delaying a choice during search is
called a least commitment strategy.
Example
• Consider the problem of putting on a pair of
shoes.
Partial-order Plan
(putting on a pair of shoes)
• Plan is a two-action sequence
– RightSock followed by RightShoe to achieve the first
conjunct of the goal
– LeftSock followed by LeftShoe for the second conjunct.
• Two sequences can be combined to yield the final plan.
• The two subsequences can be manipulated
independently, without committing to whether an
action in one sequence is before or after an action in
the other.
• A planning algorithm that can place two actions into a
plan without specifying which comes first is called a
partial-order planner.
Partial Order Plan Creation Steps
1. Start with a simple partial plan (start and goal, no
actions)
2. Expand the plan until the complete plan is
developed by applying the following plan operators
– Adding a step
– Imposing an ordering that puts one step after another
3. Gradually move from incomplete/vague plans to
complete, correct plans
4. Backtrack if a condition is unachievable or if a
conflict is irresolvable
Initial Partial Plans / Sub-Plans
Partial Order Plan Graph
Corresponding Set of Total order Plans
Components of a Partial order plan

Where, S is a step

Where, C is an action

Ex:
Example 2: Flat Tyre Problem
Partial Order Plan
Example 3
• Planning to buy milk and banana from a supermarket and a
drill from a hardware shop that sells them.
• The start state and goal state with the literals that are true at
that state are given.
• A progressive plan will move from start state to goal state by
performing a set of actions.
• A regressive plan will move from goal state to start state by
performing a set of actions
Heuristics for Partial-order Planning
• The heuristic function is used to choose which
plan to refine
• Planning graph can be used to give better
heuristic estimates
Planning Graph
• Planning graphs work only for propositional
planning problems (ones with no variables)
• A Planning graph consists of a sequence of
levels that correspond to time steps in the plan
– Level 0 is the initial state.
– Each level contains a set of literals and a set of actions.
– The literals are all those that could be true at that time
step, depending on the actions executed at preceding time
steps
– If some goal literal does not appear in the final level of the
graph, the goal is not achievable
Steps in creating Planning Graphs
• Start with initial state level So
• Follow that with action level Ao (actions whose
preconditions are satisfied in the previous level)
• Connect each action to its preconditions in So and its
effects in S1(introducing new literals into S1 that
were not in So)
Example : Eating a Cake
Actions: Bake, Eat
Initial State So
Add Actions from initial state S0 to
State S1
Add persistence actions from initial
state S0 to State S1

Persistence actions/ inactions: small squares


Add Mutex Relations to avoid conflicts
• A mutex relation (arcs) holds between two actions/
literals when one action/literal negates the effect of
another
• A mutex between two actions indicates that it is
impossible to perform these actions in parallel.
• A mutex between two literals indicates that it is
impossible to have these both literals true at this
stage.
Mutex Relations

• Persistence action Have(Cake) and Eat(Cake) are mutexes


(cannot be done together at the same level)
• Persistence action ˥Eaten(Cake) and Eat(Cake) are mutexes.
• Have(Cake) and ˥Have(Cake) are mutexes
• Eat(Cake) and ˥Eat(Cake) are mutexes
Termination
Repeat process until graph levels off
• two consecutive levels are identical
• contain the same amount of literals
Planning Graph for the
Flat Tyre Changing Problem
Heuristic Estimation using Planning
Graphs
• The cost estimate of any goal literal (level cost)
is the level at which it first appears in the
planning graph.
– Eaten( Cake) has level cost 1.
Heuristics cost for
a conjunction of goals
• The cost of a conjunction of goals is given by
1. max-level heuristic (maximum level cost of any of
the goals)
2. level sum heuristic(sum of the level costs of the
goals)
3. set-level heuristic (level at which all the literals in
the conjunctive goal appear in the planning
graph without any pair of them being mutually
exclusive)
Example
• The heuristic estimate for the conjunctive goal
Have(Cake) And Eaten (Cake) is
 2(max-level )
 0 + 1 = 1(level sum)
 2(set level)
GraphPlan Algorithm
GraphPlan Algorithm
• GRAPHPLAN alternates between a
– solution extraction step
– graph expansion step
• EXTRACT-SOLUTION looks for whether a plan can be
found starting at the end and searching backwards
• EXPAND-GRAPH adds the actions for the current
level and the state literals for the next level

You might also like