Haley: A Hierarchical Framework for Logical Composition of Web Services

21
Haley: A Hierarchical Framework for Logical Composition of Web Services Haibo Zhao , Prashant Doshi LSDIS Lab, Dept. of Computer Science, University of Georgia IEEE International Conference on Web Services 2007

description

IEEE International Conference on Web Services 2007. Haley: A Hierarchical Framework for Logical Composition of Web Services. Haibo Zhao , Prashant Doshi LSDIS Lab, Dept. of Computer Science, University of Georgia. Outline. Introduction Motivating scenario Background - PowerPoint PPT Presentation

Transcript of Haley: A Hierarchical Framework for Logical Composition of Web Services

Page 1: Haley: A Hierarchical Framework for Logical Composition of Web Services

Haley: A Hierarchical Framework for Logical

Composition of Web Services

Haibo Zhao, Prashant DoshiLSDIS Lab, Dept. of Computer Science,

University of Georgia

IEEE International Conference on Web Services 2007

Page 2: Haley: A Hierarchical Framework for Logical Composition of Web Services

Outline Introduction

Motivating scenario

Background

Model: First order Semi-Markov decision processes (FO-SMDP)

Composing nested Web processes using Haley

Architecture

Experiment & Discussion

Page 3: Haley: A Hierarchical Framework for Logical Composition of Web Services

Introduction Web service composition

Business processes with Web services as components

Existing approaches to composition: AI planning Classical planning techniques

Golog, Model checking-based planning, HTN planning, Synthy Decision-theoretic planning

MDP (Doshi2004)

Limitations Classical planning assumes deterministic behavior of Web services Guarantee correctness but not optimality State space explosion Cannot operate directly on WS descriptions in FOL

Page 4: Haley: A Hierarchical Framework for Logical Composition of Web Services

Our Approach: HaleyStochastic SMDP model

Handle uncertainties in WS invocation Provide cost-based optimality

Hierarchical model to represent the hierarchies in Web processes Address the scalability problem

FOL based representation Directly operate on WS description in FOL

Page 5: Haley: A Hierarchical Framework for Logical Composition of Web Services

Handling Orders in Supply ChainLevel 1: Composition using composite FO-SMDP

Abstract action

Level 0: Composition using primitive FO-SMDP

Level 0: Composition using primitive FO-SMDP

Abstract action

Page 6: Haley: A Hierarchical Framework for Logical Composition of Web Services

Background: Probabilistic Situation Calculus [Reiter01] A FOL based framework for representing actions, changes and

reasoning about them

Probabilistic Situation calculus elementsActions: parameterirzed FO termsReceiveOrder(o)

Situations: sequence of actions representing the state of the world do(ReceiveOrder(o),s0)

Fluents: situation-dependent relations and functions whose truth values may vary HaveOrder(o, s)

Nature’s Choices: capture stochastic results of actionsChoice(CheckCustomer(o), a) ≡ a = CheckCustomerS(o) ∨ a = CheckCustomerF(o)

Probabalities for nature’s choices:Pr(CheckCustomerS(o),CheckCustomer(o), s) = 0.9Pr(CheckCustomerF(o),CheckCustomer(o), s) = 0.1

Precondition Axioms:HaveOrder(o, s) ⇒ Poss(CheckCustomer(o), s)

Successor state Axioms: Describe the effects on fluentsPoss(a, s) ⇒HaveOrder(o, do(a, s)) ⇔ a = ReceiveOrderS(o) ∨ (HaveOrder(a, s) ∧ a ≠CancelOrderS(o))

Page 7: Haley: A Hierarchical Framework for Logical Composition of Web Services

First Order MDP(FO-MDP) [Boutilier 01]

Probabilistic situation calculus representation allows concise specification of complex domains

Specify lump sum reward/cost and utilities with case notationkCase(A(x)) = case[ A(x) = CheckCustomer(o), 2; A(x)= VerifyPayment(o), 3; A(x) = ChargeMoney(o),2 ]

Avoid explicit state and action enumeration

A decision-theoretic regression algorithm for solving FO-MDPs

Page 8: Haley: A Hierarchical Framework for Logical Composition of Web Services

First Order Semi-MDPs (FO-SMDP) FO-SMDP is a temporal generalization of FO-MDPs:

The sojourn time of actions are modeled with a density function; and the system will incur an action-duration cost at an accumulating rate

Case notation of sojourn time distribution

Case notation of accumulating rate

Total reward of a state-action pair

Representing total reward and utilities with case notation, FO-SMDP can be solved analogously to FO-MDP using DT regression

Page 9: Haley: A Hierarchical Framework for Logical Composition of Web Services

Level 1: Composition using composite FO-SMDP

Abstract action

Abstract action

Level 0: Composition using primitive FO-SMDP

Level 0: Composition using primitive FO-SMDP

Page 10: Haley: A Hierarchical Framework for Logical Composition of Web Services

Elicitation of Model Parameters (level 0)

Level 0: Model parameters may be obtained from WSDL-S/SAWSDL, OWL-S descriptions of Web services, and service level agreements

Compile situation calculus axioms from preconditions and effects

e.g. WS: ChargeMoney(o)Precondition: V alidCustomer(o) AND V alidPayment(o)Effect: Charged(o)

The precondition axiom:ValidCustomer(o, s) ∧ V alidPayment(o, s) ⇒ Poss(ChargeMoney(o), s)

The successor state axiom:Poss(a, s) ⇒ Charged(o, do(a, s)) ⇔ a=ChargeMoneyS(o)∨Charged(o, s)

Elicit non-functional parameters from service level agreement:

Page 11: Haley: A Hierarchical Framework for Logical Composition of Web Services

Deriving Model Parameters for Abstract Actions (level≥1)Level ≥1: Derive model parameters related to abstract actions from

lower level Web process

We need to know successor state axioms and the case notations of lump sum cost, sojourn time distribution and accumulating rate

Page 12: Haley: A Hierarchical Framework for Logical Composition of Web Services

Deriving Model Parameters for Abstract Actions Successor state axioms

– Let

– We have

And

– The successor state axiom of VerifyOrder(o) becomes:

Relation between high-level fluents and low-level fluents

Relation between high-level abstract actions and low-level actions

Page 13: Haley: A Hierarchical Framework for Logical Composition of Web Services

Deriving Model Parameters for Abstract Actions

Lump sum cost K – lump sum cost of the abstract action is the total of

lump sum costs of the corresponding primitive actions

– Add a new case into the case notation of K

kVO = kCase(CheckCustomer(o)) + kCase(VerifyPayment(o)) + kCase(ChargeMoney(o))

)K(a )K(a )K(a ) a K( CMVPCCvo

New case to be added

Page 14: Haley: A Hierarchical Framework for Logical Composition of Web Services

Deriving Model Parameters for Abstract Actions

Sojourn time distribution F – Assume the sojourn time of all primitive actions follows

Gaussian distribution: fCC(t)=N(t; µcc, σcc), fvo(t)=N(t; µvo, σvo) and fcm(t)=N(t; µcm, σcm)

– Linear combination of Gaussian distributions is a Gaussian distribution, the abstract action VerifyOrder also follows Gaussian fvo(t)=N(t; µvo, σvo) where:

– Add a new case into the case notation of F

cmvpccvo 222cmvpccvo

New case to be added

Page 15: Haley: A Hierarchical Framework for Logical Composition of Web Services

Deriving Model Parameters for Abstract Actions

Cost Accumulating Rate C – Accumulated cost of an abstract action is the total accumulated cost of

all corresponding primitive actions

– Add a new case into case notation of C

Given model parameters for abstract actions, composite FO-SMDP can be solved analogously to a primitive FO-SMDP

Page 16: Haley: A Hierarchical Framework for Logical Composition of Web Services

Architecture of Haley

Page 17: Haley: A Hierarchical Framework for Logical Composition of Web Services

Interleaved Generation and Execution of Nested Web Process

If the action is a primitiveaction

Yes

While the goal state can not be entailed by the KB

Query KB for the current state of the process

Get the optimal action based on the policy

Invoke the corresponding Web service

Recursively invoke a- low level BPEL process

Recursively call this algorithm

-for the low level Web process

No

Update KB with the effects

Page 18: Haley: A Hierarchical Framework for Logical Composition of Web Services

Performance Evaluation−Comparison with HTN planning and MBP planning on supply chain scenario

−Execute the processes generated by three approaches in a simulated environment 1000 times, measure average rewards

The performance of HTN approaches ours as the environment becomes less uncertain; Haley provides cost based optimization compared to MBP planner

Page 19: Haley: A Hierarchical Framework for Logical Composition of Web Services

Performance EvaluationComparisons of different decision theoretic planners in the same domain and the collected runtimes

Hierarchal decomposition significantly improves the performance

First order representation avoids the explicit state enumeration

Page 20: Haley: A Hierarchical Framework for Logical Composition of Web Services

Discussion• Many AI planning based approaches

AI classical planning is not designed to handle WS composition Assumes deterministic behavior of Web services Cannot directly operate on WS descriptions in FOL Does not scale well to large problems

• Haley: our hierarchical framework Stochastic optimization manages uncertainty and delivers

optimality Able to operate directly on WS descriptions in FOL Exploits hierarchy scalability Better performance in uncertain environments

• Future work Incorporate data mediation in Haley

Page 21: Haley: A Hierarchical Framework for Logical Composition of Web Services

Thank You!

Questions?

Contact usHaibo Zhao: [email protected]

Prashant Doshi: [email protected]