Invent the Future

Dynamic Web Based Methods and Tools forDynamic Web Based Methods and Tools forMulti-University I/UCRC Management, DataMulti-University I/UCRC Management, Data

Integration and Decision SupportIntegration and Decision Support

Janis TerpennyJanis Terpenny

January 8, 2009January 8, 2009

Invent the Future

Outline

• Project Objective• Brief Background on Center for e-Design• Project Tasks and Approach• Progress and Future Work

Invent the Future

Project Objective

Creation of a web-based environment that host methods and tools that are generalized and customizable to support the management and collaboration needs of multi-university centers

01101010100101010010111101011010101001010010100100111

0110101010010101001011100101010010101001011110001Real Time Collaboration and Sharing

01111100101101010100101010National Science Foundation Industry/University Cooperative Research Center for e-Design: IT-Enabled Design and Realization of Engineered Products and Systems

University of Pittsburgh UMassAmherst

Multi-University Center

http://e-design.iems.ucf.edu/

InformationKnowledgeDecision-MakingDistributedCollaborativeVisualizationSynthesisAnalysisEvaluationRisk/Uncertainty

Multi-UniversityFirst Rights to IP

Leading ResearchExcellent Students

ConsortiumTechnology

SuppliersConsumers

01101010100101010010111101011010101001010010100100111

0110101010010101001011100101010010101001011110001

01111100101101010100101010

Ontology Development and Utilization for Knowledge Management in Product Design

Step2: Ontology Structure Design

Step 4: Knowledge Base Development by Protégé

Step1:Classes and Slots Categorization

Step3: Relations Retrieval and Realization

Part

Length

Thickness

Part material

Welding

Type of Welding

Number of weldpasses

Filler Materialused

Type of fillermaterial

Diameter of weldwire

Unit cost of fillermaterialTotal cost of

electrode/joint

VariablePower

Weld currentassumed

Arc Voltageassumed Average unit cost

of power

Total variablepower cost/unit

ShieldingGas

Gas flow rateassumed

Unit gas cost

Shielding gascomposition

Shielding gascapacity

Total shielding gascost/joint

Labor

Time for weldpreparation

Operating factor

Unit cost of laborTotal laborcost/joint

Grey arrow are arrows in "classes and slots" view.Boxed items are classes.Font Color differentiates

different aspects. Colored arrows represent that thestart point decides the value of end point.

Relationship

Machinery

Machinerycost/joint

Total cost of thejoint

Welding location

Type of machineMachinery cost

Machine used year

Machinery unitcost

Operator type

Amps

VoltageDuty cycle

Solid wire size range

Develop ontology instances with ontology structure (step 2),relation(step 3) and form the knowledge base

Step5: Protégé Based Search and Reasoning

Search and reason related data based on step 1~4 to help designers to get and compare alternatives

Part

Length

Thickness

Part material

Welding

Type of WeldingNumber of weld

passes

Filler Materialused

Filler material type

Diameter of weldwire

Unit cost of fillermaterial

Total cost ofelectrode/joint

VariablePower

Weld currentassumed

Arc Voltageassumed Average unit cost

of power

Total variablepower cost/unit

ShieldingGas

Gas flow rateassumed

Unit gas cost

Shielding gascomposition

Shielding gascapacity

Total shielding gascost/joint

Operator

Time for weldpreparation

Operating factor

Unit cost ofoperator

Total laborcost/joint

Boxed items are classes.Font colorsdifferentiate different aspects.

Classes and Slots

Machinerycost/joint

Total cost of thejoint

Machinery

Type of machineMachinery cost

Machine used year

Machinery unitcost

Welding location

Operator type

Amps

Voltage

Duty cycle

Solid wire size range

Part 1

Part 2

labor

Shielding gas used

machine used

Filler metal

Power used

Step6: Ontology-based Decision Support Tool for Design for Manufacturing (DfM)

National Science Foundation Industry/University Cooperative Research Center for e-Design

Setup - 1....103255641 + 10325644 = sub assembly A

Setup - 2....sub assembly A + 10325647 = sub assembly B

Setup - 3....sub assembly B + 10325834 = sub assembly C ( final sub assembly )

Sub assembly steps

Criteria

Choose the best design alternative by considering the criteria concurrently

I/UCRCTIE Grant 0632758

Invent the Future

Project Tasks and Approach

Decision Support for Multi-University Centers

1. New Project Selection or Evaluation of Project for Continuation

2. Match Project and/or Company Needs with Specific Center Site(s)

Environment to Integrate Heterogeneous Distributed Resources

Invent the Future

Decision Support for Multi-University Centers

Project Evaluation

Site Center 1Site Center n

Outcome estimation

Cost EstimationOutcome estimation

Cost Estimation

… …

Invent the Future

Cost Estimation

Equipment Cost. Machine Cost. Material Cost . Power Cost

. Maintenance cost, etc.

Overhead Cost. Depreciation

cost. Legal Fees. Travel Cost

Etc.

Personnel Cost.Labor Cost

(Undergraduate /graduate students,

staffs, faculties)

Software Cost:. Numerical Modeling Software

. Simulation software

Miscellaneous

e.g., Center for Friction Stir Welding

Machine Cost CM = [ሺWT×nሻ+Ts+Tch ]×CMRMR×60

Labor Cost CL = (WT×nOF +TWP) × CLR60

Where, WT=Time to weld; n= Number of weld passes; Ts=Setup Time (min); Tch=Tool change-over time (min); MR=Machine Reliability (Assuming 95%); CMR=Machine Rate($/hr); OF=Operation Factor;

TWP=Time for weld preparation (min); CLR=Labor Rate ($/hr)

Sources of Costs

Invent the Future

Sources of Benefits

Research Contribution

Intellectual Property

Students receiving degrees

. Ph.D. Master

. Undergraduate

Publications.Journal paper. Conference

paper

Presentations

Patent Applications

Software copyrights

Patents Granted/Derived

Invention Disclosures

Royalties Realized

Liscensing Agreement

Project benefit

Centers

R&D-Increased technical

awareness,. Accelerated or

new project. Development of

Intellectual property

Professional Networking

. Improved ability to recruit students

. Increased cooperation with other industrial members and

scientists

Commercialization. Improved products

. New products. Process. Service

Sales.New or Retained

jobs

Center Members

Invent the Future

Decision Support

Two types of approaches for project selection:1) Rule-based decision making

- If (Conditions), then (action)

For example:

A class ‘Project’ will be defined with attributes, such as cost (equipment cost and Personnel cost). The projects that will be evaluated are instances of class ‘Project’.

Rule 1: If (Cost<=$15,000) &(Number of Patents>=1), then ‘Start project ’;

Rule 2: If (New product=‘available’), then ‘Start Project’

Invent the Future

2) Algorithm-based decision making Weighted Sum (weights determined by

decision makers)U(A)=w1*U(Cost)+w2*U(Project benefit)

Where,

A= Project ‘A’;

w1=Weight for criteria ‘cost’; w2=Weight for criteria ‘Project benefit’;

U(cost)=Utility value for total cost;

It is a decreasing function, the less the cost is, the larger the utility value, e.g. u(Cost) = − e –a*Cost;

U(Project benefit)=Composite utility value for the project benefit.

U(Project benefit)=U(Project contribution in Research) +U(Intellectual Property )+U(R&D)+U(Commercialization)+U(Professional networking)

Decision Support

Invent the Future

Rank and Order• Rank the criteria and order the projects

Attribute RankCost Total Cost 1Research Contribution Publications 2

Presentations 3Students Receiving degrees 4

Intellectual Pro. Invention Disclosure 5Patent Applications 6Software Copyright 7Patent Granted/Derived 8Liscencing Agreement 9Royalties Realized 10

Center Members Research & Development 11Commericialization 12Professional Networking 13

Attribute Rank Attribute Project 2 Project 3 Project 4 Project 11 Publications 4 4 3 22 Total cost ($) 55,000 45,000 60,000 65,0003 Presentation 2 3 1 44 Patent Applications 2 0 0 05 Students Receiving degrees 1 1 2 06 Invention Disclosure 0 0 0 07 Software Copyright 0 0 0 08 Patent Granted/Derived 0 0 0 09 Liscencing Agreement 0 1 1 0

10 Royalties Realized 0 0 0 011 Research & Development Good Excellent Good Excellent12 Commericialization Not bad Good Excellent Good13 Professional Networking Bad Good Excellent Good

Attribute Rank AdjustedCost Total Cost 1 2Research Contribution Publications 2 1

Presentations 3 3Students Receiving degrees 4 5

Intellectual Pro. Invention Disclosure 5 6Patent Applications 6 4Software Copyright 7 7Patent Granted/Derived 8 8Liscencing Agreement 9 9Royalties Realized 10 10

Center Members Research & Development 11 11Commericialization 12 12Professional Networking 13 13

Invent the Future

Framework - Decision Support

Data Integration

DatabaseDatabase

Existing projectsCapabilitiesFacilitiesExpertise

Modelseg. .Cost Model .Benefit Model

Data Process

Customization System

Other Systems

New Project Input

Confirmation and

Adjustment

Database

Potential Projects

Decision Making Models

eg. . Rule-based Decision . Aggregated Decision

Calculation results

Data Layer

Project Evaluation

System

User Interface

Invent the Future

Integration Environment

central ontology attribute

databaseattribute

table

attribute name

related databaseattribute table name

tablename

CentralOntol ogy

LocalOntol ogy 1

LocalOntol ogy 2

LocalOntol ogy n

Data Resource 1 Data Resource 2 Data Resource n

. . .

. . .

Web Portal

Data Query Guidance

Data Resources Specification

Querying Sentences Formation

Central Ontology Local Ontologies

Query Result Analysis

Heterogeneous Data Resources

Structure of Local Ontologies

DESIGNPROCESSOPTIMIZATIONINTEGRATION TRADES SIMULATION VISUALIZATION

www.phoenix-int.com© Copyright 2009 Phoenix Integration, Inc. All Rights Reserved

Background: Phoenix Integration

Provide engineering software and services to customers in aerospace, defense, and related industries

14 year history

Evolved out of a research program at Virginia Tech

Office locations Philadelphia, PA (Corporate) Blacksburg, VA (R&D) California (Sales) North East (Sales)

World-wide sales in North America, Europe, and Asia

www.phoenix-int.com

DESIGNPROCESSOPTIMIZATIONINTEGRATION TRADES SIMULATION VISUALIZATION

www.phoenix-int.com© Copyright 2009 Phoenix Integration, Inc. All Rights Reserved

Core Product: ModelCenter®

Visual environment for process integration

Graphically link analyses together

Automatically transfer data from analysis to analysis

Reduce data transfer errors

Save time Perform trade studies

to find better designs1.Create Models2.Generate Data3.Interpret Results

DESIGNPROCESSOPTIMIZATIONINTEGRATION TRADES SIMULATION VISUALIZATION

www.phoenix-int.com© Copyright 2009 Phoenix Integration, Inc. All Rights Reserved

CENTER CONNECTIVITY GOALS

LINKING NSF CENTERS

LINKING NSF CENTERS

LINKING CENTER SITES

DECISION SUPPORT

TECHNOLOGY READINESS

EDUCATIONAL

CONTRIBUTION ECONOMIC BENEFIT

SHARE BESTPRACTICES

REPORTING AMONGCENTERS & NSF

SYNERGY WITHINMULTI-UNIVERSITYCENTERS

OPEN ARCHITECTURE

KNOWLEDGEMINING

Invent the Future

Progress and Future Work

Working closely with CFSP to identify criteria, data sources (automated and manual) for decision support

Extending prior work in web-based ontological approach to the integration of heterogeneous distributed resources

Evaluating feasibility and trade-offs of using commercially available tools (e.g., partnering with Phoenix Integration)

Invent the Future

Example Questions

1. Factors to consider in the economic evaluation of projects?

2. How to measure these criteria: convert all criteria into $ or utility or measure with nominal scales

3. What is the commonly used method for decision making? (rule based or multi-criteria such as 'rank and order' or attribute aggregation)

4. When deciding which center site for a project, what attributes are considered (expertise, geographical proximity, etc)?

5. What other measures are used to quantify the attributes of a center site?

Invent the Future

Thanks!

Grateful Acknowledgment for funding from the National Science Foundation through Grants

EEC-0632758, EEC-0542084, IIP 0840402

Any opinions, findings, and conclusions or recommendations presented in this paper are those of the authors and do not necessarily reflect the views of the

National Science Foundation