OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Application of Artificial Intelligent Planning
byKomminist (04305201)
Somasekara Rao (05407008)
November 13, 2005
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Motivation
AI Planning is ubiquitous in on a daily basis life, for example, fromplanning how to make dinner to planning how to graduate fromUniversity with the least amount of work. Researchers in AIcommunity have studied AI planning problems for many decades,and many techniques exist for automating planning processes.This seminar explore both conventional and modern approaches toAI planning. Issues to be discussed include: how to representactions and world state, how to search for plans efficiently, andmore .. by the help of example(s)
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Outline
Introduction
Need of AI Planning
Properties of AI Planning
Types of Planning Problem and Algorithms
Planning in Application
Summary and Conclusions
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
What is AI Planning?
Definition Of AI planning
...Planning every where ...
Planning is a problem solving technique.Planning is reasoning about future events in order to verify theexistence of a reasonable series of actions to take in order toaccomplish a goal.
... we do planning ...
to reduce searching,to resolve conflicts, andto provide basis for error recovery.
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
What is AI Planning?
AI Planning ...
In more general sense : Given
a way to describe the worldan initial state of the worlda set of goals to achieveda set of possible action to change the world
FindA prescription for actions to change the initial state into onethat fulfill the goal
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
What is AI Planning?
Representation of Change ...
As actions change the world or we consider possible actions,we want to :
Know how an action change the world ,Keep track of the history of the world states (have we beenhere before?)Answer questions about potential world states (what wouldhappened if...?)
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Planning domains ...
Web management agents :
An Efficient Plan Execution System for InformationManagement Agents for gathering information from the WorldWide Web ...
Robot planning
For controlling the legs, arms and action of the robot (perhapsmost of planning research relay here!)
Manufacturing planning, Image processing management,Linguistic transportation
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Domain Cond ...
Crisis management , Bank risk management, Blocks-world,
Puzzles, Artificial Domain,
Software planning ( postponed example later ) and more ...
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Properties of Planning Algorithms
SoundnessA planning algorithm is sound if all solutions found are legalplans
All preconditions and goals are satisfiedNo constraints are violated (temporal, variable binding)
CompletenessA planning algorithm is complete if a solution can be foundwhenever one actually existsA planning algorithm is strictly complete if all solutions areincluded in the search space
OptimalityA planning algorithm is optimal if the order in which solutionsare found is consistent with some measure of plan quality
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Types of Planning
Linear planning - Planning with a goal stack.
Work on one goal until completely solved before moving on tothe next goalMore feasible if zero goal interaction !
Nonlinear planning - Handles goal interactions byInterleaving.
Use goal set instead of goal stackInclude in the search space all possible subgoal orderings
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Linear planning : discussion ...
Basic Idea
Work on one goal until completely solved before moving on tothe next goal
Planning algorithm maintains goal Stack
Implications
No interleaving of goal achievementEfficient search if goals do not interact much !!!
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Means-Ends Analysis
Basic Idea
Search only relevant aspects of problemWhat means (operators) are available to achieve the desiredends goal
Find difference between goal and current state
Find operator to reduce difference
Perform means-ends analysis on new subgoals ...
Used often in Agent-Oriented Software Modellinglanguage, such as i* , Tropos,...
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
GPS (Newell, Simon, Ernst, 1960s)
Introduced concept of means-ends analysis
Essentially linear planning using recursive procedure calls asthe goal-stack mechanism
Simple and easy as well as the first planning algorithm .
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
STRIPS (Fikes, Nilsson, 1971)
Same basic idea as GPS, but
Solved the frame problem (STRIPS assumptions)
Actions must specify all the state variables whose values theychange...No disjunction allowed in effects and like that ...
Introduced operator representation
Operationalized notion of difference, subgoals, and operatorapplication
Dealt (somewhat) with plan execution and learning.
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Pros and Cons on Linearity
Advantages
Reduced search space, since goals are solved one at a time.Advantageous if goals are (mainly) independentLinear planning is sound
Disadvantages
Linear planning may produce suboptimal solutions (based onthe number of operators in the plan)Linear planning is incomplete.
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Example : One-way Rocket
(Op :LOADprecAt(obj,loc)At(roc,loc)effectsadd : inside(obj,roc)del : At(obj,loc)
Op : UNLOADpreinside(obj,roc)At(roc,loc)effectsadd : At(obj,loc)del : inside(obj, roc)
Op : MOVEpreAt(roc,from)has-fuel(roc)effectsadd : At(roc,to)del : At(roc,from),del : has-fuel(roc)
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Incompleteness of Linear Planning
Initial state :At(obj1,locA)At(obj2, locA)At(roc, locA)has-fuel(roc)
Goal Statement:( and At(obj1,locB)At(obj2, locB )
Goal Plan
At(obj1,locB) LOAD-ROCKET obj1 locA)(MOVE-ROCKET)(UNLOAD-ROCKET obj1 locb)
At(obj2,locB)failure
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
State-Space Search
Node denoted a state of the world.
Arcs corresponds to the executionof a specific action
Planning problem : finding a path.
Difficulty : requires CompleteDescription of the states searched,and requires the search to becarried out locally
Can be progression or regression ...
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Plan-space search
Each node in the graph representspartial plans.
Arcs denote plan refinementoperations.
One can search for a plan : totalorder planning, or partial orderplanning.
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Partial Order Planning
Basic Idea :
Search in plan space and use least commitment, when possible
Initial Plan :
< {start, finish}, {start < finish, {}}start has preconditions; Its effects are the initial statefinish has no effects; Its preconditions are the goals
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
POP contd ...
A set of actions : For example,{eat-breakfast (go to mess) , take-shower, wake-up, go-to-AISeminar }
A set of ordering constraints : For example,{wake-up before eat-breakfast, wake-up before take-shower,wake-up before go-to-AI Seminar, take-shower beforego-to-work}
A set of causal links ( set of bindings ) : For example,wake-up −−−awake−−− >eat-breakfastis a link from the action, wake-up to the action eat-breakfast.Note : When the action wake-up is added to the plan, theabove causal link is recorded, along with the orderingconstraint {wake-up before eat-breakfast}
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Least Commitment
Basic IdeaMake choices only that are relevant to solving the current partof the problem
Least commitment choicesOrderings : Leave actions unordered, unless they must besequentialBindings : Leave variables unbounded, unless needed to unifywith conditions being achieved.Actions : Usually not subjected to ”least commitment”
RefinementOnly add information to the current planTransformational planning can remove choices
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Many more non-linear planning
Graph-based searchSat-based searchOBDD-based search
Hierarchical planningEmphasis on action decomposition/refinementKnowledge engineering/acquisitionVery little search
many more ...
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Classical Deterministic planning
State of the world at any moment is unambiguouslydetermined by the initial state of the world and sequence ofactions that have been taken.
Action Model : Complete, deterministic, correct, richrepresentation
State : single initial state, fully known
Goals : complete satisfaction
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Deterministic planning contd ...
Given set of states = black dots,actions = (blue, red) and initialstate = I and a set of goal states= G
The task is to find the a pathfrom initial state to one of thegoal states........
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Non-deterministic Planning
The world is not always deterministic in nature !!! Why ...?
At least for two reasons:The model of the World is usually very incomplete !
We do not know whether somebody is going to phone or visitus, and then the visit or phone call can be modeled as anondeterministic event that may or may not take place.
Second, many actions themselves are by their naturenondeterministic, either intentionally or unintentionally.
Throwing two dice has 12 possible outcomes that usuallycannot be predicted (which is why throwing dice isinteresting!)Throwing some object to a garbage bin from a distance mayor may not succeed Try it after this !!!
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Types of planningLinear PlanningPlanning as SearchDeterministic vs Non deterministic
Non-deterministic contd ...
we added a red arrow to thesecond state above the initialstate,
as well as a new state to thebottom of right corner with anondeterministic transitionleading to it.
the new arrows make the redaction nondeterministic, and animpact on which action sequencesare plans.
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
State - Space Nonlinear Planning
Extend linear planningUse goal set instead of goal stackInclude in the search space all possible subgoal orderingsHandles goal interactions by interleaving
AdvantagesNon-linear planning is soundNon-linear planning is completeNon-linear planning may be optimal with respect to planlength (depending on search strategy employed)
DisadvantagesLarger search space, since all possible goal orderings may haveto be consideredSomewhat more complex algorithm; More bookkeeping
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Completeness of State-Space Nonlinear Planning
State-space nonlinear planning is Complete( Continued from the incompleteness of the linear planning ...)
Goal Plan
At(obj1,locB) (LOAD obj1 locA)
At(obj2,locB) (LOAD obj1 locA)
At(obj1,locB) (MOVE)(ULOAD obj1 locB)
At(obj2,locB) (ULOAD obj2 locB)
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Generation a Solution plan
A complex process :
Alternative operators to achieve a goal.Multiple goals that interactsEfficiency and plan quality.
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
More Planning under nonlinear
NOAM [Sacerdoti1975]
First nonlinear, partial-order plannerIntroduce the notion of plan-space searchUsed TOME (Table Of Multiple Effects) to detect goalinteractions
SNLP (Systematic Non Linear Planner), UCPOP (UniversalConditional POP )
SHOP2 ( Simple Hierarchical Order Planner )
SAVTA (Subgoal After Every Try to Apply), SABA ( SubgoalAlways Before Applying)
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Applying AI Planning for SEAI Planning for Tool construction
Planning for AOSE : Motivation to i*
i* : graphical modeling language used in AOSE (tropos)
i* => a framework for modeling social settings, based on thenotions of actors,goals, dependencies ( social interactions )...
i* framework consists of two graphical models: SD and SR
The SD diagram is used to represents the strategicdependencies.
Which express intentional relationships that exist amongactors in order to fulfill some strategic objectives.
Dependency type : goal Dependency, Softgoal Dependency,Task Dependency,Resource Dependency
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Applying AI Planning for SEAI Planning for Tool construction
SD : for the CAS
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Applying AI Planning for SEAI Planning for Tool construction
SR Model: b/n Student and RAO
SR provides a more detailed level of modeling than SD bylooking in-side actor.
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Applying AI Planning for SEAI Planning for Tool construction
So where is AI planning used?
Goal decomposition/refinment ( Conjuction and /ordisjuction) ...
GetAdmission : (Apply, PassExam, GetAdmResult)
Ordering (partial or total) should has to consider (prior-to):For example
Apply is possible iff you have pass exam !!
Imagin that there is high goal interaction among actors viadependencies !!
To get admission a student needs admission co-ordinator whichis the RAO
So here is the challenge!! But carefully choose your planner!
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Applying AI Planning for SEAI Planning for Tool construction
Example-I
Initial State Set Of Objects
part1, type: PART
drill-1, type: SPOT-DRILL
drill-2, type: TWIST-DRILL
Goal Statement
has-hole(part)by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Applying AI Planning for SEAI Planning for Tool construction
Example-I Contd..
drill-spot(part, drill-bit)Pre: (holding-tool drill-bit)
(holding-part part)Add: (has-spot part)
drill-hole(part, drill-bit)Pre: (has-spot part)
(holding-tool drill-bit)(holding-part part)
Add: (has-hole part)
put-drill-bit(drill-bit)Pre: (tool-holder-empty)Add: (holding-tool drill-bit)Del: tool-holder-empty
remove-drill-bit(drill-bit)Pre: (holding-tool drill-bit)Add: tool-holder-emptyDel: (holding-tool drill-bit)
put-part(part)Pre: part-holder-emptyAdd: (holding-part part)Del: part-holder-empty
remove-part(part)Pre: (holding-part part)Add: part-holder-emptyDel: (holding-part part)by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Applying AI Planning for SEAI Planning for Tool construction
Example-I contd... sequence of actions
put-part(part)put-drill-bit(drill-1)drill-spot(part drill-1)remove-drill-bit(drill-1)put-drill-bit(drill-2)drill-hole(part drill2)
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Applications : Current and Potential
Scheduling problems with action choices as well as resourcehandling requirements
Problems in supply chain managementHSTS (Hubble Space Telescope scheduler)Workflow management
Autonomous agentsRAX/PS (The NASA Deep Space planning agent)
Software module integratorsTest case generation (Pittsburgh)
Interactive decision supportMonitoring subgoal interactions
Plan-based interfaces
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Current research direction
Planning application in Robotics
Planning for web-service composition
Planning databasesDistributed Intelligent Agents Group: For example
Multiagent Planning and Coordination as Distributed SearchMultiagent Planning and Coordination for Unmanned GroundVehiclesIntelligent Agent Infrastructures for Supporting Collaborationmore description =>http://ai.eecs.umich.edu/diag/homepage.html
AI Planning in Grid Computing
AI planning in Tropos Methodology and more .....
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Conclusion drawn
No general planning algorithm
Tackling real-world planning problems often requiresconsidering various types of constraints, ranging from simplenumerical comparators to complex resources.
There is a method how to solve planning tasks within generalconstraint-solving frameworks, such as propositionalsatisfiability, integer programming, and constraintprogramming. ”paper : Constraints and AI Planning”
Planning as search : control rules to capture heuristics forefficient search; learning opportunities!
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Conclusion contd ...
Applying AI planning to Early-phase of SE ensures quality ofSoftware Development in both functional and non-functionalaspect of the system.
For example, Tropos Software engineering use planning forsearching of interactive goal in the specification to handleconsistencies among different actors hence which is moreamendable for formal verification and specification ofsoftwares.
Every body has to practice planning to become rich ! (Please do planning now onwards!!)
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
References and Links
AIPS 2002 Sixth International conference on ArtificialIntelligence Planning and Scheduling (AIPS2002) Workshopon Knowledge Engineering Tools and Techniques for AIPlanning Toulouse, France 24 April 2002
Alexander Nareyek, Eugene C. Freuder, Robert Fourer, EnricoGiunchiglia, Robert P. Goldman, Henry Kautz, Jussi Rintanen,Austin Tate. ”Constraints and AI Planning,” IEEE IntelligentSystems, vol. 20, no. 2, pp. 62-72, March/April, 2005.
Biplav Srivastava and Jana Koehler : Planning withWorkflows - An Emerging Paradigm for Web ServiceComposition, ECOOP December 2004.
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
References and Links Contd ...
J. Koehler, J. Hoffmann: “On Reasonable and Forced GoalOrderings and their Use in an Agenda-Driven PlanningAlgorithm”, Journal of Artificial Intelligence Research,12/2000, pages 339-386.
B. Nebel, Y. Dimopoulos, J. Koehler,“ Ignoring IrrelevantFacts and Operators in Plan Generation”, ECP-97, pages338-350
“The Role of Planning in Grid Computing” Jim Blythe, EwaDeelman, Yolanda Gil, Carl Kesselman, Amit Agarwal,Gaurang Mehta and Karan Vahi , International Conference onAutomated Planning and Scheduling, ICAPS 2003
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
References and Links Contd ...
Finding plans and defining plan constraints simultaneously”Jim Blythe KCAP 01 Workshop on Interactive Tools forKnowledge Capture, 2001
Thomas A. Montgomery and Edmund H. Durfee, “SearchReduction in Hierarchical Distributed Problem Solving”.Group Decision and Negotiation 2:301-317 (Special issue onDistributed Artificial Intelligence), 1993.
http://www.cs.cmu.edu/ reids/planning/notes.html ( Date :10 / 10 / 2005 )
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
References and Links Contd ...
http://www.laas.fr/planning/ ( date : 4/11/2005)
http://www.planet-noe.org (Date: 23/10/2005 )
http://www.informatik.uni-freiburg.de/ ki/lehre/ss04/aip/(Date 10/10/2005)
Book: Artificial Intelligence A modern approach”, Russel. Sand Norvig, Pearson Education Publishers, 2003.
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning
OutlineIntroduction
Where AI Planning usedProperties of AI Planning
Types of Planning Problems and AlgorithmsNon-Linear Planning
Application of AI PlanningCurrent research in AI Planning
Conclusion and SummaryReferences
Thank You
by Komminist (04305201) Somasekara Rao (05407008) Application of Artificial Intelligent Planning