Administración de Proyectos

Importance of Project Management

• Projects represent change and allow organizations to effectively introduce new products, new process, new programs

• Project management offers a means for dealing with dramatically reduced product cycle times

• Projects are becoming globalized making them more difficult to manage without a formal methodology

• Project management helps cross-functional teams to be more effective

Management of IT Projects

• More than $250 billion is spent in the US each year on approximately 175,000 information technology

projects.

• Only 26 percent of these projects are completed on time and within budget.

• The average cost for a development project for a large company is more than $2 million.

• Project management is an $850 million industry and is expected to grow by as much as 20 percent per year.

Bounds, Gene. “The Last Word on Project Management” IIE Solutions, November, 1998.

What Defines a Project?

•••••••

How does a project differ from a

program?

Project Management versus Process Management

“Ultimately, the parallels between process and project management give way to a fundamental difference: process management seeks to eliminate variability whereas project management must accept variability because each project is unique.”

Elton, J. & J. Roe. “Bringing Discipline to Project Management” Harvard

Business Review, March-April, 1998.

Measures of Project Success

•••••••

Was the movie “Titanic” a success?

Delayed Openings are a Fact of Life in the Foodservice, Hospitality Industry

Disney's shipbuilder was six months late in delivering its new cruise ships, and thousands of customers who had purchased tickets were stranded. Even with that experience, their second ship was also delivered well after the published schedules. Universal Studios in Orlando, Fla. had been building a new restaurant and entertainment complex for more than two years. They advertised a December opening, only to announce in late November that it would be two or three months late.

Even when facilities do open close to schedule, they are rarely finished completely and are often missing key components. Why do those things happen? With all of the sophisticated computers and project management software, why aren't projects completed on schedule?

Frable, F. Nation's Restaurant News (April 12, 1999)

IT Project Outcomes

More than 200% late

Cancelled

On-Time

Less than 20% late

21-50% late

51-100% late

101-200% late

26%

29%

6%

16%

9%

8%

6%

Source: Standish Group Survey, 1999 (from a survey of 800 business systems projects)

Why do Projects Fail?Studies have shown that the following factors contribute significantly to project failure:

• Improper focus of the project management system

• Fixation on first estimates

• Wrong level of detail

• Lack of understanding about project management tools; too much reliance on project management software

• Too many people

• Poor communication

• Rewarding the wrong actions

Why do IT Projects Fail?

• Ill-defined or changing requirements

• Poor project planning/management

• Uncontrolled quality problems

• Unrealistic expectations/inaccurate estimates

• Naive adoption of new technology

Source: S. McConnell, Construx Software Builders, Inc.

QuickTime™ and aPhoto - JPEG decompressor

are needed to see this picture.

Have You Ever Lost Sight of the Project Goals?

Not all Projects Are Alike…

“[in IT projects], if you ask people what’s done and what remains to be done there is nothing to see. In an IT project, you go from zero to 100 percent in the last second--unlike building a brick wall where you can see when you’re halfway done. We’ve moved from physical to non-physical deliverables….”

J. Vowler (March, 2001)

Engineering projects = task-centric

IT projects = resource-centric

Shenhar’s Taxonomy of Project Types

Degree ofUncertainty/Risk

System Complexity/Scope

High

Low-Tech

AssemblyProjects

ArrayProjects

SystemProjects

Medium-Tech

High-Tech

Super High-Tech

Construction

Newcellphone

New shrink-wrappedsoftware

ERPimplementationin multi-national

firm

Auto repair

Advancedradarsystem

Project Life Cycle

TimePhase 1 Phase 2 Phase 3 Phase 4Formation & Planning Scheduling & Evaluation &Selection Control Termination

Req

uire

d R

esou

rces

Life Cycle Models: Pure WaterfallConcept Design

Requirements Analysis

Architecture Design

Detailed Design

Coding & Debugging

System Testing

Source: S. McConnellRapid Development (Microsoft Press, 1996)

Life Cycle Models: Code & Fix

Design, Cost, Time Trade-offs

Target

COST

DE

SIG

N

TIME (SCHEDULE)

Due Date

Budget Constraint

Optimal Time-Cost Trade-off

Required Performance

Optional Scope Contracts

Fixed Scope Contract specifies SCHEDULE, COST, SCOPE

Optional Scope Contract specifies SCHEDULE, COST, QUALITY

(general design guidelines may be indicated)

Since it is widely accepted that you can select three of the four dimensions (or perhaps only

two), what to do?

Importance of Project Selection

“There are two ways for a business to succeed at new products: doing projects right, and doing the right projects.”

Cooper, R.G., S. Edgett, & E. Kleinschmidt. Research • Technology Management, March-April, 2000.

Project Initiation & Selection

• Critical factors 1) Competitive necessity 2) Market expansion 3) Operating requirement

• Numerical Methods 1) Payback period 2) Net present value (NPV) or Discounted Cash Flow (DCF) 3) Internal rate of return (IRR) 4) Expected commercial value (ECV)

• Project Portfolio 1) Diversify portfolio to minimize risk 2) Cash flow considerations 3) Resource constraints

Payback Period

Number of years needed for project to repay its initial fixed investment

Example: Project costs $100,000 and is expected to save company $20,000 per year

Payback Period = $100,000 / $20,000 = 5 years

Net Present Value (NPV) Discounted Cash Flow (DCF)

Let Ft = net cash flow in period t (t = 0, 1,..., T) F0 = initial cash investment in time t = 0

r = discount rate of return (hurdle rate)

NPV = Ft

1 + r tt = 0

T

Internal Rate of Return (IRR)

Find value of r such that NPV is equal to 0

F0 + F11 + r

+ F2

1 + r 2 = 0

Example (with T = 2):

Find r such that

DCF Project Example*

*Hodder, J. and H.E. Riggs. “Pitfalls in Evaluating Risky Projects”, Harvard Business Review, Jan-Feb, 1985, pp. 128-136.

Product Demand Product Life

Annual Net Cash Inflow Probability

High 20 years $24 million 0.3Medium 10 years $12 million 0.5

Low Abandon Project None 0.2

Phase I Research and Product Development$18 million annual research cost for 2 years60% probability of success

Phase II Market DevelopmentUndertaken only if product development is successful$10 million annual expenditure for 2 years to develop marketing anddistribution channels (net of any revenues earned in test marketing)

Phase III SalesProceeds only if Phase I and II verify opportunity.Production is subcontracted and all cash flows are after-tax and occur at year's end.The results of Phase II (available at the end of year 4) identify the product's market potential as indicated below:

DCF Project Example (cont’d)

Year Expected Cash Flow (in $ million)1 -182 -183 0.6 (-10) = - 64 0.6 (-10) = - 6

5 - 14 .6 (0.3 x 24 + 0.5 x 12) = 7.9215 - 24 .6 (0.3 x 24) = 4.32

What is the internal rate of return for this project?

DCF Example Continued

What if you can sell the product (assuming that both Research and Product Development AND Market Development are successful) to a third party? What are the risks AT THAT POINT IN TIME?

Assume that discount rate r2 is 5%

ProbabilityWhat is 20 years of cash inflow at $24M/year? $299.09 0.3What is 10 years of cash inflow at $12M/year? $92.66 0.5

Expected value of product at Year 4: $136.06

DCF Example Continued

Expected cash flows (with sale of product at end of year 4) are now:

Outflow Inflow Net ProbabilityExpected

Cash FlowYear 1 18.00$ (18.00)$ 1 (18.00)$ Year 2 18.00$ (18.00)$ 1 (18.00)$ Year 3 10.00$ (10.00)$ 0.6 (6.00)$ Year 4 10.00$ 136.06$ 126.06$ 0.6 75.63$

What is the internal rate of return for this project?

Criticisms of NPV/DCF1) Assumes that cash flow forecasts are accurate; ignores

the “human bias” effect

2) Fails to include effects of inflation in long term projects

3) Ignores interaction with other proposed and ongoing projects (minimize risk through diversification)

4) Use of a single discount rate for the entire project (risk is typically reduced as the project evolves)

Expected Commercial Value (ECV)

Develop New Product

Technical Failure

Technical Success

Probability = pt

Probability = 1 - pt

Launch New Product

Commercial Failure (with net

benefit = 0)

Commercial Success (with net benefit =

NPV)

Probability = pc

Probability = 1 - pc

Risk class 1 Risk class 2

DCF Example Revisited

Discount rate r1 Discount rate r2

Research & Product

Development

Development Succeeds

Probability = pt

Development Fails

Probability = 1 - pt

Market Development

Product Demand High

0.3

Product Demand Medium

Product Demand Low

0.5

0.2

Drop project

Ranking/Scoring ModelsProfitability/value1) Increase in profitability?2) Increase in market share?3) Will add knowledge to organization that can be leveraged by other projects?4) Estimated NPV, ECV, etc.

Organizat ion's Strategy1) Consistent with organization's mission statement?2) Impact on customers?

Risk1) Probability of research being successful?2) Probability of development being successful?3) Probability of process success?4) Probability of commercial success?5) Overall risk of project6) Adequate market demand?7) Competitors in market

Organizat ion Costs1) Is new facility needed?2) Can use current personnel?3) External consultants needed?4) New hires needed?

Miscellaneous Factors1) Impact on environmental standards?2) Impact on workforce safety?3) Impact on quality?4) Social/political implications

Scoring Attributes

vi xi = 1 - exp L - xi

1 - expL - U .

To convert various measurement scales to a (0, 1) range….

LINEAR SCALE: value of attribute i is

EXPONENTIAL SCALE: value of attribute i is

vi xi = xi - LU - L

0.000.10

0.200.30

0.400.50

0.600.70

0.800.90

1.00

1 2 3 4 5 6 7Response

Att

ribu

te V

alue

Linear ScaleExponential Scale

Ranking/Scoring Example

Attribute Measurement ScaleAttribute

Weight (wi)1) Does project increase market share? unlikely likely 30%2) Is new facility needed? yes no 15%3) Are there safety concerns? likely unsure no 10%4) Likelihood of successful technical development? unlikely likely 20%

5) Likelihood of successful commercial development? unlikely likely 25%

1 2 3 4 5

1 2 3 4 5

1 2 3 4 5

Attribute #1 #2 #3 #4 #5Project

Score (V j)Project A 4 yes likely 4 1Project B 2 no unsure 3 4

Linear ScaleProject A 0.75 0.25 0 0.75 0 0.413Project B 0.25 0.75 0.5 0.5 0.75 0.525

Exponential ScaleProject A 0.97 0.64 0.00 0.97 0.00 0.581Project B 0.64 0.97 0.88 0.88 0.97 0.845

Ranking/Scoring Example (cont’d)

Analyzing Project Portfolios: Bubble Diagram

Expected NPV

Prob of Commercial Success

HighZero

Low

High

Analyzing Project Portfolios: Product vs Process

Ext

ent o

f Pro

duct

Cha

nge

Extent of Process Change

Source: Clark and Wheelwright, 1992

Key Elements of Project Portfolio Selection Problem

1. Multi-period investment problem

2. Top management typically allocates funds to different product lines (e.g., compact cars, high-end sedans)

3. Product lines sell in separate (but not necessarily independent) market segments

4. Product line allocations are changed frequently

5. Conditions in each market segment are uncertain from period to period due to competition and changing customer preferences

“Stage-Gate” Approach

Installation PlanFacility PrepTraining Plan Implementation

Detail DesignSchedule & BudgetContingency PlanProduct & Performance Reviews

Initiation Define Design ControlImprove

Work StatementRisk AssessmentPurchasing PlanChange Mgt

InitiationProject ReviewCharter

Source: PACCAR Information Technology DivisionRenton, WA

Production close-out

Lessons learned

Post-project audit

Project Selection Example

Y e a r (t)1 2 3 4

Project A ($40) $10 $20 $20Project B ($65) ($25) $50 $50

Budget Limit (B t) $120 $20 $40 $55

Phases of Project Managementn Project formulation and selectionn Project planning

u Summary statementu Work breakdown structureu Organization planu risk managementu Subcontracting and bidding process

n Project schedulingu Time and scheduleu Project budgetu Resource allocationu Equipment and material purchases

n Monitoring and controlu Cost control metricsu Change ordersu Milestone reports

Project Planningn Summary Statement

u Executive summary: mission and goals, constraintsu Description and specifications of deliverablesu Quality standards used (e.g., ISO)u Role of main contractor and subcontractorsu Composition and responsibilities of project team

n Organization Planu Managerial responsibilities assigned; signature authority u Cross impact matrix (who works on what)u Relationship with functional departmentsu Project administrationu Role of consultantsu Communication procedures with organization, client, etc.

Importance of Project Planning

The 6P Rule of Project Management:

Prior Planning Prevents Poor Project Performance

“If you fail to plan, you will plan to fail”

Anonymous

Work Breakdown Structure (WBS)

1) Specify the end-item “deliverables”

2) Subdivide the work, reducing the dollars and complexity with each additional subdivision

3) Stop dividing when the tasks are manageable “work packages” based on the following:

• Skill group(s) involved• Managerial responsibility• Length of time• Value of task

Work Packages/Task Definition

The work packages (tasks or activities) that are defined by the WBS must be:

• Manageable

• Independent

• Integratable

• Measurable

Design of a WBS

“The usual mistake PMs make is to lay out too many tasks; subdividing the major achievements into smaller and smaller subtasks until the work breakdown structure (WBS) is a ‘to do’ list of one-hour chores. It’s easy to get caught up in the idea that a project plan should detail everything everybody is going to do on the project. This springs from the screwy logic that a project manager’s job is to walk around with a checklist of 17,432 items and tick each item off as people complete them….”

The Hampton Group (1996)

Two-Level WBS

1. Charity Auction

1.1 Event Planning

1.2 Item Procurement

1.3 Marketing 1.4. Corporate Sponsorships

WBS level 1

WBS level 2

Three-Level WBS

1.1 Event Planning

1.2 Item Procurement

1.3 Marketing

1. Charity Auction

1.4 Corporate Sponsorships

1.1.1 Hire Auctioneer

1.1.2. Rent space

1.1.3 Arrange for decorations

1.2.1 Silent auction items

1.2.2 Live auction items

1.2.3 Raffle items

1.3.1 Individual ticket sales

1.3.2 Advertising

1.1.4 Print catalog

WBS level 1

WBS level 2

WBS level 3

Estimating Task Durations (cont’d)

• Benchmarking• Modular approach• Parametric techniques• Learning effects

Beta Distribution

Completion time of task j

Optimistic Time tjo Pessimistic Time tj

p

Time

Probability density function

Expected duration = Most Likely Time = tm

Beta DistributionFor each task j, we must make three estimates:

most optimistic timemost pessimistic timemost likely time

tjo

tjp

tjm

Expected duration j = tjo + tj

p + 4tjm

6

Variance of task j = j2 =

tjp - tj

o 2 36

Estimating Task Durations: Painting a RoomTask: Paint 4 rooms, each is approximately 10’ x 20’. Use flat paint on walls, semi-gloss paint on trim and woodwork. Each room has two doors and four windows. You must apply masking tape before painting woodwork around the doors and windows. Preparation consists of washing all walls and woodwork (some sanding and other prep work will be needed). Only one coat of paint is necessary to cover existing paint. All supplies will be provided at the start of the task. Previous times on similar painting jobs are indicated in the table below.

hours min hours min27 25 31 5238 25 19 1533 12 26 2717 44 30 2726 7 25 2122 1 24 2814 2 32 5830 27 32 128 30 13 4321 13 42 4523 59 22 5727 44 32 1523 15 32 3137 6 27 1517 54 26 1117 13 21 52

What is your estimate of the average time you will need? What is your estimate of the variance?

Estimating Task Durations with Incentives

Task: Consider the painting job that you have just estimated. Now, however, there are explicit incentives for meeting your estimated times. If you finish painting the room before your specified time, you will receive a $10 bonus payment. HOWEVER, if you finish the painting job after your specified time, you will be fined $1000.

Revised estimated time =

Estimating Task Durations with Incentives

Task: Consider the painting job that you have just estimated. Now, however, there are explicit incentives for meeting your estimated times. If you finish painting the room before your specified time, you will receive a $10 bonus payment. If you finish the painting job after your specified time, there is no penalty.

Revised estimated time =

Role of Project Manager/Team

Project Manager

Client

Subcontractors

Regulating Organizations

Project Team

Functional Managers

Top Management

Responsibilities of a Project ManagerTo the organization and top management • Meet budget and resource constraints • Engage functional managers

To the project team • Provide timely and accurate feedback • Keep focus on project goals • Manage personnel changes

To the client • Communicate in timely and accurate manner • Provide information and control on changes/modifications • Maintain quality standards

To the subcontractors • Provide information on overall project status

Project Team

What is a project team?A group of people committed to achieve a common set of goals for which they hold themselves mutually accountable

Characteristics of a project team• Diverse backgrounds/skills• Able to work together effectively/develop synergy• Usually small number of people• Have sense of accountability as a unit

“I design user interfaces to please an audience of one. I write them for me. If I’m happy, I know some cool people will like it. Designing user interfaces by committee does not work very well; they need to be coherent. As for schedule, I’m not interested in schedules; did anyone care when War and Peace came out?”

Developer, Microsoft CorporationAs reported by MacCormack and Herman, HBR Case 9-600-097: Microsoft Office 2000

Intra-team CommunicationM = Number of project team members

L = Number of links between pairs of team members

If M =2, then L = 1

If M =3, then L = 3

Number of Intra-team Links

0

50

100

150

200

250

300

350

400

450

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

N = No. of Team Members

L =

No.

of I

ntra

-tea

m L

inks

L = Number of Intra-team Links = N2

= N(N-1)2

Importance of Communication

On the occasion of a migration from the east, men discovered a plain in the land of Shinar, and … said to one another, “Come, let us build ourselves a city with a tower whose top shall reach the heavens….” The Lord said, …“Come, let us go down, and there make such a babble of their language that they will not understand one another’s speech.” Thus, the Lord dispersed them from there all over the earth, so that they had to stop building the city.

Genesis 11: 1-8

Project Performance and Group Harmony

Two schools of thought:

1) “Humanistic school” -- groups that have positive characteristics will perform well

2) “Task oriented” school -- positive group characteristics detract from group performance

What is the relationship between the design of multidisciplinary project teams and project success?

Project Performance and Group Harmony (cont’d)

Experiment conducted using MBA students at UW and Seattle U using computer based simulation of pre-operational testing phase of nuclear power plant*

Total of 14 project teams (2 - 4 person project teams) with a total of 44 team members; compared high performance (low cost) teams vs low performance (high cost) teams

Measured: Group HarmonyGroup Decision Making EffectivenessExtent of Individual’s Contributions to GroupIndividual Attributes

*Brown, K., T.D. Klastorin, & J. Valluzzi. “Project Management Performance: A Comparison of Team Characteristics”, IEEE Transactions on Engineering Management, Vol 37, No. 2 (May, 1990), pp. 117-125.

Group Harmony: High vs Low Performing Groups

4.004.204.404.604.805.005.205.405.605.806.00

1 2 3 4 5 6 7Week

Gro

up H

arm

ony

High Performance (low cost) Teams Low Performance (high cost) Teams

Extent of Individual Contribution: High vs Low Performing Groups

4.004.204.404.604.805.005.205.405.605.806.00

1 2 3 4 5 6 7Week

Exte

nt o

f Ind

ivid

ual C

ontr

ibut

ions


Decision Making Effectiveness: High vs Low Performing Groups

3.00

3.50

4.00

4.50

5.00

5.50

6.00

1 2 3 4 5 6 7Week

Dec

isio

n M

akin

g Eff

ecti

vene

ss


Project Organization Types• Functional: Project is divided and assigned to appropriate functional

entities with the coordination of the project being carried out by functional and high-level managers

• Functional matrix: Person is designated to oversee the project across different functional areas

• Balanced matrix: Person is assigned to oversee the project and interacts on equal basis with functional managers

• Project matrix: A manager is assigned to oversee the project and is responsible for the completion of the project

• Project team: A manager is put in charge of a core group of personnel from several functional areas who are assigned to the project on a full-time basis

Project Organization Continuum

Project Team Organization

Project Matrix

Project fully managed by functional managers Project fully managed by

project team manager

FunctionalOrganization

Functional Matrix

Balanced Matrix

A Business School as a Matrix Organization

Dean

Associate Dean for Undergraduate

Program

Associate Dean for MBA Programs

Director of Doctoral Program

Accounting Department Chair

Marketing Department Chair

Finance Department Chair

Gloria

Diane

Bob

ZeldaLarry

Curly

Moe

Barby

Leslie

Matrix Organizations & Project Success

• Matrix organizations emerged in 1960’s as an alternative to traditional means of project teams

• Became popular in 1970’s and early 1980’s

• Still in use but have evolved into many different forms

• Basic question: Does organizational structure impact probability of project success?

Organizational Structure & Project Success• Studies by Larson and Gobeli (1988, 1989)

• Sent questionnaires to 855 randomly selected PMI members

• Asked about organizational structure (which one best describes the primary structure used to complete the project)

• Perceptual measures of project success: successful, marginal, unsuccessful with respect to :

1) Meeting schedule2) Controlling cost3) Technical performance4) Overall performance

• Respondents were asked to indicate the extent to which they agreed with each of the following statements:

1) Project objectives were clearly defined2) Project was complex3) Project required no new technologies4) Project had high priority within organization

• Classification of 547 respondents (64% response rate)30% project managers or directors of project mgt programs16% top management (president, vice president, etc.)26% managers in functional areas (e.g., marketing)18% specialists working on projects

• Industries included in studies14% pharmaceutical products10% aerospace10% computer and data processing products

others: telecommunications, medical instruments, glass products, software development, petrochemical products, houseware goods

• Organizational structures:13% (71): Functional organizations26% (142): Functional matrix16.5% (90): Balanced matrix28.5% (156): Project matrix16% (87): Project team

Study Data

ANOVA Results by Organizational Structure

Controlling Cost

Meeting Schedule

Technical Performance

Overall Results

Organizational Structure N Ave (SD) Ave (SD) Ave (SD) Ave (SD)

AFunctional

Organization 71 1.76 (.83) 1.77 (.83) 2.30 (.77) 1.96 (.84)

B Functional Matrix 142 1.91 (.77) 2.00 (.85) 2.37 (.73) 2.21 (.75)

C Balanced Matrix 90 2.39 (.73) 2.15 (.82) 2.64 (.61) 2.52 (.61)

D Project Matrix 156 2.64 (.76) 2.30 (.79) 2.67 (.57) 2.54 (.66)

E Project Team 87 2.22 (.82) 2.32 (.80) 2.64 (.61) 2.52 (.70)

Total Sample 546 2.12 (.79) 2.14 (.83) 2.53 (.66) 2.38 (.70)

F-statistic 10.38* 6.94* 7.42* 11.45*

Scheffe ResultsA,B < C,D,E

E < D A,B < C < D,E A,B < C,D,E A,B < C,D,E

*Statistically significant at a p<0.01 level

Summary of Results• Project structure significantly related to project success

• New development projects that used traditional functional organization had lowest level of success in controlling cost, meeting schedule, achieving technical performance, and overall results

• Projects using either a functional organization or a functional matrix had a significantly lower success rate than the other three structures

• Projects using either a project matrix or a project team were more successful in meeting their schedules than the balanced matrix

• Project matrix was better able to control costs than project team

• Overall, the most successful projects used a balanced matrix, project team, or--especially--project matrix

Subcontracting = Business Alliancen When you subcontract part (or all) of a

project, you are forming a business alliance....

Intelligent Business Alliances: “A business relationship for mutual benefit between two or more parties with compatible

or complementary business interests and/or goals”

Larraine Segil, Lared Presentations

Communication and Subcontractors

How is knowledge transferred?

What types of communication mechanism(s) will be used between company and subcontractor(s)?

WHAT a company communicates.....

HOW a company communicates.....

Personality Compatibility

Corporate Personality

Subcontractor Personality

Individual Personality

Project

Subcontracting Issues

n• What part of project will be subcontracted?n• What type of bidding process will be used? What type of

contract?n• Should you use a separate RFB (Request for Bids) for

each task or use one RFB for all tasks? n• What is the impact on expected duration of project?n• Use a pre-qualification list?n• Incentives? Bonus for finishing early? Penalties for

finishing after stated due date?

• What is impact of risk on expected project cost?

Basic Contract Typesn Fixed Price Contract

u Client pays a fixed price to the contractor irrespective of actual audited cost of project

n Cost Plus Contract u Client reimburses contractor for all audited costs of project (labor, plant,

& materials) plus additional fee (that may be fixed sum or percent of costs incurred)

n Units Contractu Client commits to a fixed price for a pre-specified unit of work; final

payment is based on number of units produced

Incentive (Risk Sharing) Contracts

General Form:Payment to Subcontractor = Fixed Fee + (1 - B) (Project Cost)

where B = cost sharing rate

Cost Plus Contract

B = 0 B = 1

Fixed Price Contract

Linear & Signalling Contracts

Why Use Incentive Contracts?Expected Cost of Project = $100M

Two firms bid on subcontract

Firm 1 Firm 2

Fixed Fee (bid) $5 M $7 M

Project Cost $105 M $95 M

(inefficient producer)

What is result if Cost Plus Contract (B = 0) used?

Washington State Bid Code (WAC 236-48-093)n WAC 236-48-093: A contract shall be awarded to the lowest responsible and responsive

bidder based upon, but not limited to, the following criteria where applicable and only that which can be reasonably determined:

n 1) The price and effect of term discounts...price may be determined by life cycle costing if so indicated in the invitation to bid

n 2) The conformity of the goods and/or services bid with invitation for bid or request for quotation specifications depicting the quality and the purposes for which they are required.

n 3) The ability, capacity, and skill of the bidder to perform the contract or provide the services required.

n 4) The character, integrity, reputation, judgement, experience, and efficiency of the bidder.

n 5) Whether the bidder can perform the contract with the time specified.n 6) The quality of performance on previous contracts for purchased goods or services.n 7) The previous and existing compliance by the bidder with the laws relating to the

contract for goods and services.n 8) Servicing resources, capability, and capacity.

Competitive Bidding: Low-Bid System

n “In the low-bid system, the owner wants the most building for the least money, while the contractor wants the least building for the most money. The two sides are in basic conflict.”

Steven GoldblattDepartment of Building ConstructionUniversity of WashingtonThe Seattle Times, Nov 1, 1987

Precedence Networks

Networks represent immediate precedence relationships among tasks (also known as work packages or activities) and milestones identified by the WBS

Milestones (tasks that take no time and cost $0 but indicate significant events in the life of the project)

Two types of networks: Activity-on-Node (AON)

Activity-on-Arc (AOA)

All networks: must have only one (1) starting and one (1) ending point

Precedence Networks: Activity-on-Node (AON)

A

B

C

D

Start End

Precedence DiagrammingStandard precedence network (either AOA or AON) assumes that a successor task cannot start until the predecessor(s) task(s) have been completed. Alternative relationships can be specified in many software packages:

Finish-to-start (FS = ): Job B cannot start until days after Job A is finished

Start-to-start (SS = ): Job B cannot start until days after Job A has started

Finish-to-finish (FF = ): Job B cannot finish until days after Job A is finished

Start-to-finish (SF = ): Job B cannot finish until days after Job A has started

Critical Path Method (CPM): Basic Concepts

Task A7 months

Task B3 months

End

Task C

11 months

Start

Critical Path Method (CPM): Basic Concepts

Start

Task A7 months

Task B3 months

Task C11 months

End

ESStart = 0LFStart = 0

ESA = 0LFA = 8

ESB = 7LFB = 11

ESC = 0LFC = 11

ESEnd = 11LFEnd = 11

ESj = Earliest starting time for task (milestone) j

LFj = Latest finish time for task (milestone) j

AON Precedence Network: Microsoft Project

Start

1 0dWed 12/20/00 Wed 12/20/00

Task A

2 7dWed 12/20/00 Thu 12/28/00

Task C

4 11dWed 12/20/00 Wed 1/3/01

End

5 0dWed 1/3/01 Wed 1/3/01

Task B

3 3dFri 12/29/00 Tue 1/2/01

Critical Path Method (CPM): Example 2

Task A 14 wks

Task D 12 wks

Task E 6

wks

Task B 9 wks

Task C 20 wks

Task F 9

wks

START END

ESF =LFF =

ESD =LFD =

ES START = 0LFSTART = 0

ESA =LFA =

ESB =LFB =

ESEND =LFEND =

ESC =LFC =

ESE =LFE =

Example 2: Network Paths

Path TasksExpected

Duration (wks)1 START-A-D-F-END 352 START-A-D-E-END 323 START-B-D-F-END 304 START-B-D-E-END 275 START-C-E-END 26

Example 2: CPM Calculations

E A R L I E S T L A T E S T

Task or Milestone

Duration ( )

Start Time (ESi) Finish Time Start Time

Finish Time (LFi)

START 0 0 0 0 0A 14 0 14 0 14B 9 0 9 5 14C 20 0 20 9 29D 12 14 26 14 26E 6 26 32 29 35F 9 26 35 26 35

END 0 35 35 35 35

ti

Example 2: Calculating Total Slack (TSi)

Task or Milestone

Duration ( )

Earliest Start Time

(ESi)

Lastest Finish Time

(LFi)Total Slack

(TSi)Critical Task?

START 0 0 0 0 YesA 14 0 14 0 YesB 9 0 14 5 NoC 20 0 29 9 NoD 12 14 26 0 YesE 6 26 35 3 NoF 9 26 35 0 Yes

END 0 35 35 0 Yes

ti

Total Slack for task i = TSi = LFi - ESi - ti

Slack (Float) Definitions (for task i)

Total Slack (TSi) = LFi - ESi - ti

Free Slack (FSi) = ESi,min - ESi - ti

where ESi,min = minimum early start time of all tasks that

immediately follow task i

= min (ESj for all task j Si)

Safety Slack (SSi) = LFi - LFi,max - ti

where LFi,max = maximum late finish time of all tasks that

immediately precede task i

= min (LFj for all task j Pi)

Independent Slack (ISi) = max (0, ESi,min - LFi,max - ti)

Example #2: LP ModelDecision variables: STARTj = start time for task j

END = ending time of project (END milestone)

Minimize END subject to

STARTj ≥ FINISHi for all tasks i that immediately precede task j

STARTj ≥ 0 for all tasks j in project

where FINISHi = STARTi + ti = STARTi + duration of task i

Example #2: Excel Solver Model

Gantt Chart

ID Task Name1 Start

2 Task A

3 Task B

4 Task C

5 Task D

6 Task E

7 Task F

8 Task G

9 Task H

10 Task J

11 End

3/1

Workers[5]

Workers[7]

Workers[3]

Workers[12]

Workers[2]

Workers

Workers[2]

Workers[5]

Workers[6]

5/10

21 24 27 1 4 7 10 13 16 19 22 25 28 31 3 6 9 12 15 18 21 24 27 30 3 6 9 12 15 18February March April May

Microsoft Project 4.0

Project Budgeting• The budget is the link between the functional units and the project

• Should be presented in terms of measurable outputs

• Budgeted tasks should relate to work packages in WBS and organizational units responsible for their execution

• Should clearly indicate project milestones

• Establishes goals, schedules, and assigns resources (workers, organizational units, etc.)

• Should be viewed as a communication device

• Serves as a baseline for progress monitoring & control

• Update on rolling horizon basis

• May be prepared for different levels of aggregation (strategic, tactical, short-range)

Project Budgeting (cont’d)

• Top-down Budgeting: Aggregate measures (cost, time) given by top management based on strategic goals and constraints

• Bottom-up Budgeting: Specific measures aggregated up from WBS tasks/costs and subcontractors

Issues in Project Budgets• How to include risk and uncertainty factors?

• How to measure the quality of a project budget?

• How often to update budget?

• Other issues?

Critical Path Method (CPM): Example 2

Task A 14 wks

Task D 12 wks

Task E 6

wks

Task B 9 wks

Task C 20 wks

Task F 9

wks

START END

ESF = 26LFF = 35

ESD = 14LFD = 26

ES START = 0LFSTART = 0

ESA = 0LFA = 14

ESB = 0LFB = 14

ESEND = 35LFEND = 35

ESC = 0LFC = 29

ESE = 26LFE = 35

Project Budget Example

Task or Milestone

Duration (tj)

Early Start Time (ESj)

Latest Start Time (LSj)

No. of Resource A

workers

No. of Resource B

workersMaterial

CostsDirect Labor

Cost/wkLabor +

MaterialsSTART 0 0 0 - - - - -

A 14 0 0 2 0 340$ 800$ 1,140$ B 9 0 5 4 12 125$ 8,800$ 8,925$ C 20 0 9 3 14 -$ 9,600$ 9,600$ D 12 14 14 0 8 200$ 4,800$ 5,000$ E 6 26 29 1 0 560$ 400$ 960$ F 9 26 26 4 10 90$ 7,600$ 7,690$

END 0 35 35 - - - - -

Cost for Resource A worker = $400/week

Cost for Resource B worker = $600/week

Project Budget Example (cont’d)Early Start Times

Task 1 2 3 4 5 6 7 8 9 10 11 12A 1140 800 800 800 800 800 800 800 800 800 800 800B 8925 8800 8800 8800 8800 8800 8800 8800 8800C 9600 9600 9600 9600 9600 9600 9600 9600 9600 9600 9600 9600DEF

Weekly Subtotals 19665 19200 19200 19200 19200 19200 19200 19200 19200 10400 10400 10400Cumulative 19665 38865 58065 77265 96465 115665 134865 154065 173265 183665 194065 204465

Late Start TimesTask 1 2 3 4 5 6 7 8 9 10 11 12

A 1140 800 800 800 800 800 800 800 800 800 800 800B 8925 8800 8800 8800 8800 8800 8800 8800C 9600 9600 9600 9600DEF

Weekly Subtotals 1140 800 800 800 9725 9600 9600 9600 19200 19200 19200 19200Cumulative 1140 1940 2740 3540 13265 22865 32465 42065 61265 80465 99665 118865

W e e k

W e e k

Cumulative Costs

0

50000

100000

150000

200000

250000

300000

350000

400000

450000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33

Week

Cum

ulat

ive

Cost

Early Start Schedule Late Start Schedule

Range of feasible budgets

Weekly Costs (Cash Flows)

0

5000

10000

15000

20000

25000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33

Week

Wee

kly

Cost

s

Early Start Schedule Late Start Schedule

Managing Cash Flows

• Want to manage payments and receipts

• Must deal with budget constraints on project and organization requirements (e.g., payback period)

• Organization profitability

Cash Flow Example

M1

END

START

Task B 8 mos

Receive payment of $3000

Receive payment of $3000

Make payment of $5000

Task C 4 mos

Task A 2 mos

M2

Task D 8 mos

Task E 3 mos

Cash Flow Example: Solver Model

Material Management Issues

When to order materials? How much to order?

Example:

• Single material needed for Task B (2 units) and Task E (30 units)

• Fixed cost to place order = S

• Cost of holding raw materials proportional to number of unit-weeks in stock

• Cost of holding finished product greater than the cost of holding raw materials

• Project can be delayed (beyond 17 weeks) at cost of $P per week

Material Management Example

Task A 4 wks

Task B 8 wks

Task C 5 wks

Task D 6 wks

Task E 2 wks

Task F 3 wks

EndStart 2 units

30 units

LSA = 0 LSB = 4 LSC = 12

LSD = 6 LSE = 12 LSF = 14

Lot-Sizing Decisions in Projects• To minimize holding costs, only place orders at Late Starting Times

• Can never reduce holding costs by delaying project

Time

1 2 3 4 5 6 7 8 9 10 11 12

Demand: 2 30

Order option #1: 32

Order option #2: 2 30

Choose the option that minimizes inventory cost = order cost + holding cost of raw materials

Time-Cost Tradeoffs

Time-Cost Tradeoff Example

TaskNormal

Duration Normal Cost

Marginal Cost to Crash One

WeekA 7 $60 $8B 6 $85 $5C 15 $55 $10D 10 $120 $4

A

B

C

D

Start End

Time-Cost Tradeoff Example (cont’d)Project

Duration (weeks) Critical Path(s) Task(s) Reduced

Total Direct Cost

22 Start-A-C-End - $320

21 Start-A-C-End A $328Start-A-B-End

20 Start-A-C-End C $338Start-A-B-End

19 Start-A-C-End C $348Start-A-B-End

18 Start-A-C-End A, B $361Start-A-B-End

Linear Time-Cost TradeoffIn theory, the normal or expected duration of a task can be reduced by assigning additional resources to the task

Time

Cost

Crash Point

Normal Point

Slope (bj) = Increase in cost by reducing task by one time unit

Normal time =Crash time =

Normal cost =

Crash cost =

tjNtjc

Cjc

CjN

Balancing Overhead & Direct Costs

Project Duration

Cost

Indirect (overhead)

Costs

Direct Costs

Total Cost

Crash Time

Normal TimeMinimum Cost Solution

Time-Cost Tradeoff (Direct Costs Only)Given Normal point with cost and time

and Crash point with cost and time

Assume constant marginal cost of crashing task j =

Decision Variables: Sj = Starting time of task j

END = End time of project

tj = Duration of task j

Minimize Total Direct Cost =

Sj ≥ Si + ti for all tasks i Pj

for all tasks in project

END = Tmax

tj, Sj ≥ 0

CjN

Cjc tj

c

tjN

bj = Cj

c - CjN

tjc - tj

N

bj tjj

tjc Š tj Š tj

N

General Time-Cost Tradeoffs

where

I = indirect (overhead) cost/time period

P = penalty cost/time period if END is delayed beyonddeadline Tmax

L = number of time periods project is delayed beyond deadline Tmax

Minimize Total Costs = + I (END) + P L

bj tjj

QUESTION: HOW TO DEFINE L?

Software Project Schedules

“Observe that for the programmer, as for the chef, the urgency of the patron may govern the scheduled completion of the task, but it cannot govern the actual completion. An omelet, promised in ten minutes, may appear to be progressing nicely. But when it has not set in ten minutes, the customer has two choices--wait or eat it raw. Software customers have the same choices.

The cook has another choice; he can turn up the heat. The result is often an omelet nothing can save--burned in one part, raw in another.”

F.P. Brooks, “The Mythical Man-Month”, Datamation, Vol 20, No 12 (Dec, 1974),

pp. 44-52.

Coordination Costs (Software Development Project)n Assume you want to develop program that will require (approximately) 50,000 lines of

PERL coden A typical programmer can write approximately 1500 lines of code per weekn Coordination time is M (M-1)/2 weeks

No. of Programmers

No. of Weeks Coding

No. of Coordination

Weeks

Total Number of

Weeks1 33.33 0 33.332 16.67 1 17.673 11.11 3 14.114 8.33 6 14.335 6.67 10 16.676 5.56 15 20.567 4.76 21 25.768 4.17 28 32.179 3.70 36 39.70

10 3.33 45 48.3311 3.03 55 58.03

0

2

4

6

8

10

12

0 10 20 30 40 50 60 70Total Number of Weeks

No. o

f Pro

gram

mer

s (C

ost)

Brook’s Law

“Adding manpower to a late software project makes it later.”

n

F.P. Brooks, “The Mythical Man-Month”, Datamation, Vol 20, No 12 (Dec, 1974), pp. 44-52.

Compressing New Product Development Projects

Traditional Method

Design follows a sequential pattern where information about the new product is slowly accumulated in consecutive stages

Stage 0 Stage 1 Stage N

New Product Development Process

Overlapped Product DesignAllows downstream design stages to start before preceding upstream stages have finalized their specifications….

Stage 0

Stage 1

Stage N

Issues and Tradeoffs

What are the tradeoffs when moving from a traditional sequential product design process to an overlapped product design process?

• Increased uncertainty (that leads to additional work)

• Can add additional resources to tasks to reduce duration--but costs are increased

Classic PERT Model Defined

• Since task durations are now random variables, time of any milestone (e.g., end of project) is now RV

• Assume all tasks are statistically independent

• Use values of j to identify expected critical path

• Since time of event (e.g., ESk) is now sum of independent RV’s, central limit theorem specifies that ESk is approximately normally distributed with mean E[ESk] and variance Var[ESk]

where there exists s paths to task k

Expected early start time of task k = EESk = maxs

jtasks j on path s

Classic PERT Model (cont’d)

Expect Project Duration = E[ESEND] = jtasks j on CP

Variance of Project Duration = Var[ESEND] = j2

tasks j on CP

Thus, expected project duration is defined as:

Using central limit theorem and standard normal distribution:

P ESEND Š Tmax = P z Š Tmax - E ESEND

Var ESEND

PERT Example #1

Duration Estimates ExpectedTask Description Predecessors Optimistic Pessimistic Likely Duration Variance

A Requirements Analysis none 2 14 6 6.67 4.00B Programming A 4 12 7 7.33 1.78C Hardware acquisition A 2 13 8 7.83 3.36D User training A 12 18 14 14.33 1.00E Implementation B, C 3 7 5 5.00 0.44F Testing E 3 7 4 4.33 0.44

END End of project D, F 0 0 0 0.00 0.00

StartTask A


Task C Hardware

Acquisition

Task B Programming

Task F Testing

Task D User

Training

Task E Implementation

End

PERT Example #1 (cont’d)

ExpectedTask Path Early Start Variance Due Date Zi Pr(zi)B,C,D Start-A 6.67 4.00 6 -0.33 0.37

E Start-A-C 14.50 7.36 15 0.18 0.57F Start-A-C-E 19.50 7.81 20 0.18 0.57

End Start-A-C-E-F-End 23.83 8.25 25 0.41 0.66

PERT Expected Duration = 23.83 Expected CP = {Start, A, C, E, F, End}PERT Variance = 8.250

StartTask A


Task C Hardware

Acquisition

Task B Programming

Task F Testing

Task D User

Training

Task E Implementation

End

PERT Example #2

Task B

B = 12

B2 = 4

Task D

D = 3

D2 = 1

Task A

A = 4

A2 = 2

Task C

C = 10

C2 = 5

ENDSTART

Example #3: Discrete Probabilities

Task A Task B Task C Task DValue Prob Value Prob Value Prob Value Prob

7 0.333 2 0.2 5 0.2 3 0.38 0.333 12 0.8 15 0.2 12 0.79 0.333 25 0.6

START END

Task A(8.0)

Task B(10.0)

Task C(19.0)

Task D(9.3)

Example #3 (cont’d) Task A Task B Task C Task D Critical Prob of Length PATHS

Combination Value Prob Value Prob Value Prob Value Prob Path CP of CP A,D B, D C1 7 0.333 2 0.2 5 0.2 3 0.3 A, D 0.004 10 0.004 0.000 0.0002 7 0.333 2 0.2 5 0.2 12 0.7 A, D 0.009 19 0.009 0.000 0.0003 7 0.333 2 0.2 15 0.2 3 0.3 C 0.004 15 0.000 0.000 0.0044 7 0.333 2 0.2 15 0.2 12 0.7 A, D 0.009 19 0.009 0.000 0.0005 7 0.333 2 0.2 25 0.6 3 0.3 C 0.012 25 0.000 0.000 0.0126 7 0.333 2 0.2 25 0.6 12 0.7 C 0.028 25 0.000 0.000 0.0287 7 0.333 12 0.8 5 0.2 3 0.3 B, D 0.016 15 0.000 0.016 0.0008 7 0.333 12 0.8 5 0.2 12 0.7 B, D 0.037 24 0.000 0.037 0.0009 7 0.333 12 0.8 15 0.2 3 0.3 B, D 0.016 15 0.000 0.016 0.000

10 7 0.333 12 0.8 15 0.2 12 0.7 B, D 0.037 24 0.000 0.037 0.00011 7 0.333 12 0.8 25 0.6 3 0.3 C 0.048 25 0.000 0.000 0.04812 7 0.333 12 0.8 25 0.6 12 0.7 C 0.112 25 0.000 0.000 0.11213 8 0.333 2 0.2 5 0.2 3 0.3 A, D 0.004 11 0.004 0.000 0.00014 8 0.333 2 0.2 5 0.2 12 0.7 A, D 0.009 20 0.009 0.000 0.00015 8 0.333 2 0.2 15 0.2 3 0.3 C 0.004 15 0.000 0.000 0.00416 8 0.333 2 0.2 15 0.2 12 0.7 A, D 0.009 20 0.009 0.000 0.00017 8 0.333 2 0.2 25 0.6 3 0.3 C 0.012 25 0.000 0.000 0.01218 8 0.333 2 0.2 25 0.6 12 0.7 C 0.028 25 0.000 0.000 0.02819 8 0.333 12 0.8 5 0.2 3 0.3 B, D 0.016 15 0.000 0.016 0.00020 8 0.333 12 0.8 5 0.2 12 0.7 B, D 0.037 24 0.000 0.037 0.00021 8 0.333 12 0.8 15 0.2 3 0.3 B, D 0.016 15 0.000 0.016 0.00022 8 0.333 12 0.8 15 0.2 12 0.7 B, D 0.037 24 0.000 0.037 0.00023 8 0.333 12 0.8 25 0.6 3 0.3 C 0.048 25 0.000 0.000 0.04824 8 0.333 12 0.8 25 0.6 12 0.7 C 0.112 25 0.000 0.000 0.11225 9 0.333 2 0.2 5 0.2 3 0.3 A, D 0.004 12 0.004 0.000 0.00026 9 0.333 2 0.2 5 0.2 12 0.7 A, D 0.009 21 0.009 0.000 0.00027 9 0.333 2 0.2 15 0.2 3 0.3 C 0.004 15 0.000 0.000 0.00428 9 0.333 2 0.2 15 0.2 12 0.7 A, D 0.009 21 0.009 0.000 0.00029 9 0.333 2 0.2 25 0.6 3 0.3 C 0.012 25 0.000 0.000 0.01230 9 0.333 2 0.2 25 0.6 12 0.7 C 0.028 25 0.000 0.000 0.02831 9 0.333 12 0.8 5 0.2 3 0.3 B, D 0.016 15 0.000 0.016 0.00032 9 0.333 12 0.8 5 0.2 12 0.7 B, D 0.037 24 0.000 0.037 0.00033 9 0.333 12 0.8 15 0.2 3 0.3 B, D 0.016 15 0.000 0.016 0.00034 9 0.333 12 0.8 15 0.2 12 0.7 B, D 0.037 24 0.000 0.037 0.00035 9 0.333 12 0.8 25 0.6 3 0.3 C 0.048 25 0.000 0.000 0.04836 9 0.333 12 0.8 25 0.6 12 0.7 C 0.112 25 0.000 0.000 0.112

6.8% 32.0% 61.1%

Example #3 (cont’d)

Length of CumulativeCP's Prob Prob10 0.004 0.0011 0.004 0.0112 0.004 0.0115 0.108 0.1219 0.019 0.1420 0.019 0.1621 0.019 0.1824 0.224 0.4025 0.599 1.00

Task A Task B Task C Task D6.8% 32.0% 61.1% 38.8%

Criticality Indices

Expected Project Duration = 23.22

Monte-Carlo Simulation (PERT Example 1)Task Duration Early Latest Total Expected

Task (Uniform Dist) Start Finish Slack Duration VarianceA 4.99 0 4.99 0.00 6.67 4.00B 4.75 4.99 9.74 0.00 7.33 1.78C 3.38 4.99 9.74 1.36 7.83 3.36D 12.20 4.99 21.02 3.83 14.33 1.00E 5.94 9.74 15.68 0.00 5.00 0.44F 5.34 15.68 21.02 0.00 4.33 0.44

END 0.00 21.02 21.02 0.00 0.00 0.00

Run Project Duration t(B) t(C) t(D) t(E) t(F)1 31.07 1 0 0 1 12 27.41 0 1 0 1 13 23.97 1 0 0 1 14 28.93 0 1 0 1 15 26.85 1 0 0 1 16 28.82 0 0 1 0 07 28.77 0 1 0 1 1

197 30.37 0 1 0 1 1198 29.78 1 0 0 1 1199 25.33 1 0 0 1 1200 29.70 0 1 0 1 1Ave 27.13 48.5% 42.0% 9.5% 90.5% 90.5%Var 16.777

Project Makespan Lower Limit Upper Limit95% Confidence interval 26.56 27.7299% Confidence interval 26.37 27.90

Calculating Confidence Intervals

For a confidence interval, we can use the sample mean and the estimated standard error of the mean

where s is the sample standard deviation and n is the number of trials

Using a normal approximation, a (1- ) two-sided confidence interval is given by

sX = sn

X

X -+ z/2 sX

New Product Development Projects

START

Lease Mfg/Office

Space

Identify/hire staff

Design of physical unit

Electronics design Software

Assemble prototype Beta test prototype

END

Beta test fails (with probability of 0.25)

and rework is needed

Beta test fails (with probability of 0.25)

and rework is needed

New Product Development Projects (cont’d)

START

Lease Mfg/Office

Space

Identify/hire staff

Design of physical unit

Electronics design Software

Assemble prototype Beta test prototype

END

Beta test fails and rework is needed

Prob = .25

Prob = .75

Critical Chain and the Theory of Constraints (TOC)

• Use deterministic CPM model with buffers to deal with any uncertainties,

• Place project buffer after last task to protect the customer’scompletion schedule,

• Exploit constraining resources (make certain that resources are fully utilized),

• Avoid wasting time slack time by encouraging early task completions,

• Carefully monitor the status of the buffer(s) and communicate this status to other project team members on a regular basis, and

• Make certain that the project team is 100 percent focused on critical chain tasks

Project “Goal” (according to Goldratt): Meet Project Due Date

Project Buffer Defined

• Project Buffer is placed at the end of the project to protect the customer’s promised due date

PERT Example #1 Revisited with Project Buffer

Start

Task BProgramming

User Task D

Usertraining

Task EImplementation

End Project Buffer

Task Arequirements

analysis

Task CHardwareacquisition

Task FTesting

Calculating Project Buffer Size

For tasks k on critical chain, we can calculate project buffer using following formula that project will be completed within worst-case duration estimates around 90 percent of the time:

For those “who want a scientific approach to sizing buffers....”

Buffer = tasks k on critical chain

tkp - k

2

Implications of Project Uncertainty

Assume that the duration of both tasks A and B are described by a normal distribution with a mean of 30 days

START END

Task A

Task B

What is the probability that the project will be completed within 30 days?

Uncertainty and Worker Behavior

Consider a project with two tasks that must be completed serially

The duration of each task is described by a RV with values Ti (i = 1, 2)

Values of T 1 Prob Values of T 2 Prob7 0.3 14 0.58 0.4 18 0.59 0.3

8.0 16

Start Task 1 Task 2 End

Parkinson’s Law (Expanding Work)“Work expands so as to fill the time available for its completion”

Professor C.N. Parkinson (1957)

Set a deadline D = 24 days

So T(D) = project makespan (function of D) where

E[T(D)] = E(T1) + E(T2) + E[max(0, D - T1 - T2)]

Values of T 1 Prob Values of T 2 ProbProject

Makespan Prob7 0.3 14 0.5 24 0.157 0.3 18 0.5 25 0.158 0.4 14 0.5 24 0.28 0.4 18 0.5 26 0.29 0.3 14 0.5 24 0.159 0.3 18 0.5 27 0.15

E[T(D)] = 25 days

Procrastinating Worker

Set a deadline D = 24 days

E’[T(D)] = E(T1) + E(T2) + E{max[0, D - T1 - E(T2)]}

Can show that E[T(D)] ≥ E’[T(D)] ≥ D

What are the implications for project managers?

Values of T 1 ProbE[Delay] =

max[0, D - T1 - E(T2)] E[Makespan]

7 0.3 1 248 0.4 0 249 0.3 0 25

8 0.3 24.30

Schoenberger’s Hypothesis

An increase in the variability of task durations will increase the expected project duration….

Schoenberger’s Hypothesis Illustrated

START END

Task A

Task B

Duration of Task A Probability

Duration of Task B Probability

12 0.1 10 0.514 0.8 15 0.516 0.1

14.0 12.5

Schoenberger’s Hypothesis Illustrated

RealizationTask A

DurationTask B

Duration Probability Max (A, B)1 12 10 0.05 122 14 10 0.4 143 16 10 0.05 164 12 15 0.05 155 14 15 0.4 156 16 15 0.05 16

Duration of Task A Probability

Duration of Task B Probability

12 0.3 10 0.514 0.4 15 0.516 0.3

14.0 12.5

Increasing the variance of Task A:

Results in an increased expected duration = 14.65 days

Expected duration equals 14.55 days

Risk Management

• All projects involve some degree of risk

• Need to identify all possible risks and outcomes

• Need to identify person(s) responsible for managing project risks

• Identify actions to reduce likelihood that adverse events will occur

Risk AnalysisRisk Exposure (RE) or Risk Impact =

(Probability of unexpected loss) x (size of loss)

Example: Additional features required by clientLoss: 3 weeksProbability: 20 percentRisk Exposure = (.20) (3 weeks) = .6 week

How to Manage Project Risks?

Preventive Actions• Actions taken in anticipation of adverse events• May require action before project actually begins• Examples?

Contingency Planning• What will you do if an adverse event does occur?• “Trigger point” invokes contingency plan• Frequently requires additional costs

Risk and Contracts

High Low Low High

Degree of RiskContractor Client

Fixed Price Contract Cost Plus Contract

Firm price

Elements can be

renegotiated IncentivesT&M

with limits

Cost Plus with

IncentivesTime &

materials

Tornado Diagram

Wage Rate

Direct Labor Hours

Material Units Needed

Early Completion Bonus

Material Unit Cost

Interest rates

Energy costs

Overhead

Project Cost ($000's)

$1290

$1265

$1260

$13101310$1350

$1350

$1380$1400

$1700$1720

$1680

$1690

$1640

$1620$1625

$17601760

$1500 $1600 $1800$1700$1400$1300$1200

Sensitivity Chart

Wage Rate 0.85Direct Labor Hours 0.73

Material Units Needed 0.62Early Completion Bonus -0.45

Material Unit Cost 0.42Interest rates 0.28Energy costs 0.19

Overhead 0.10

0 0.5 1.0-0.5

Rank Order Correlation with Total Project Cost

Van Allen Company

TaskImmediate

Predecessors Time Cost Time CostStart - 0 - 0 -

A Start 3 $60 5 $40B Start 1 $50 5 $30C B 5 $70 10 $40D A 2 $60 7 $40E A 2 $50 6 $30F C, D 5 $90 11 $60G C, D 4 $60 6 $30H G 1 $40 5 $20I E, G 1 $50 4 $20

End F, H, I 0 - 0 -

Strike Expected(wks) Prob Duration

3 0.45 1.354 0.3 1.205 0.25 1.25

E[Strike Duration] 3.80

Resource Allocation & Leveling

Resource Leveling: Reschedule the noncritical tasks to smooth resource requirements

Resource Allocation: Minimize project duration to meet resource availability constraints

Resource Allocation & Leveling

Three types of resources:1) Renewable resources: “renew” themselves

at the beginning of each time period (e.g., workers)

2) Non-Renewable resources: can be used at any rate but constraint on total number available

3) Doubly constrained resources: both renewable and non-renewable

Resource Leveling

Task B2 wks

Task E3 wks

Task C9 wks

Task D5 wks

Task F2 wks

Task A3 wks

START Task G5 wks END

Task Workers Duration (t j) Early Start Late StartA 7 3 0 0B 3 2 0 3C 2 9 3 4D 10 5 3 3E 4 3 2 5F 5 2 2 11G 6 5 8 8

Resource Leveling: Early Start Schedule

Early Start Schedule

0

5

10

15

20

25

1 2 3 4 5 6 7 8 9 10 11 12 13Week

Num

ber

of W

orke

rs N

eede

d

Task GTask FTask ETask DTask CTask BTask A

Resource Leveling: Late Start Schedule

Late Start Schedule

02468

1012141618

1 2 3 4 5 6 7 8 9 10 11 12 13Week

Num

ber

of W

orke

rs N

eede

d

Task GTask FTask ETask DTask CTask BTask A

Resource Leveling: Microsoft Project

5

10

15

20

25

Workers Overallocated: Allocated:

T W T F S S M T W T F S S M T W T F S S M T W T F SDec 17, '00 Dec 24, '00 Dec 31, '00 Jan 7, '01

10 10 10 10 10 10 10 10 10 10 16 16 16 16 16 21 21 21

Renewable Resource Allocation Example (Single Resource Type)

Task B 3 wks

Task D 5 wks

Task A 4 wks

Task E 4

wksSTART

END

Task C 1 wk

3 workers

5 workers

6 workers

8 workers

7 workers

Maximum number of workers available = R = 9 workers

Resource Allocation Example: Early Start Schedule


Start End

Week 1 2 3 4 5 6 7 8 9 10 11 12No. of Workers/wk 8 8 8 11 14 8 8 8 7 7 7 7

Cumulative Workers 8 16 24 35 49 57 65 73 80 87 94 101"Wasted" worker-wks 1 1 1 - - - - - - - - -

Task B: 5 workers

Task A: 3 workers

Task C: 6 workers

Task E: 7 workersTask D:

8 workers

Resource Allocation Example: Late Start Schedule


Start End

Week 1 2 3 4 5 6 7 8 9 10 11 12No. of Workers/wk 5 5 5 11 11 11 11 14 7 7 7 7

Cumulative Workers 5 10 15 26 37 48 59 73 80 87 94 101"Wasted" worker-wks - - - - - - - - 2 2 2 2

Task B: 5 workers

Task A: 3 workers

Task C: 6 workers

Task E: 7 workersTask D:

8 workers

Resource Allocation Heuristicsn Some heuristics for assigning priorities to available tasks j, where denotes the

number of units of resource k used by task j

n 1) FCFS: Choose first available taskn 2) GRU: (Greatest) resource utilization = n 3) GRD: (Greatest) resource utilization x task duration = n 4) ROT: (Greatest) resource utilization/task duration =n 5) MTS: (Greatest) number of total successorsn 6) SPT: Shortest processing time = min {tj}n 7) MINSLK: Minimum (total) slackn 8) LFS: Minimum (total) slack per successorn 9) ACTIMj: (Greatest) time from start of task j to end of project = CP - LSj

n 10) ACTRESj: (max) (ACTIMj)n 11) GENRESj: w ACTIMj + (1-w) ACTRESj where 0 ≤ w ≤ 1

Rjk

Rjk

k

Rjk / tj

k

Rjk

k tj

Resource Allocation Problem #2

Task A1 6 days

Task A24 days

EndStart Task B1 3 days

Task C1 2 days

Task B2 5 days

Task C2 5 days

Purple Crew Gold Crew

How to schedule tasks to minimize project makespan?

Priority scheme: schedule tasks using total slack (i.e., tasks with smaller total slack have higher priority)

Task A1 Task B1 Task C11 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Task A2 Task B2 Task C21 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Gold Crew

Purple Crew

Resource Allocation Example (cont’d)

But, can we do better? Is there a better priority scheme?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Gold Crew

Purple Crew

Microsoft Project Solution (Resource Leveling Option)

Solution by: Microsoft Project 2000

Critical Chain Project Management

• Identify the critical chain: set of tasks that determine the overall duration of the project

• Use deterministic CPM model with buffers to deal with uncertainty

• Remove padding from activity estimates (otherwise, slack will be wasted). Estimate task durations at median.

• Place project buffer after last task to protect customer’s completion schedule

• Exploit constraining resource(s)

• Avoid wasting slack times by encouraging early task completions

• Have project team focus 100% effort on critical tasks

• Work to your plan and avoid tampering

• Carefully monitor and communicate buffer status

Critical Chain Buffers

Project Buffer: placed after last task in project to protect schedule

Feeding Buffers: placed between a noncritical task and a critical task

when the noncritical task is an immediate predecessor of the critical task

Resource Buffers: placed just before a critical task that uses a new

resource type

Critical Chain Illustrated

Task A1 6 days

Task A24 days

EndStart

Task B1 3 days

Task C1 2 days

Task B2 5 days

Task C2 5 days

Resource Buffers

Feeding Buffers

Non-Renewable Resources

Task B5 wks

Task D 2 wks

Task C 3 wks

Task A 6 wksSTART END

6 units

12 units

10 units

8 units

Task Duration

No. of Nonrenewable Resources Units

Needed Early Start Late StartA 6 6 0 0B 5 12 6 6C 3 10 6 8D 2 8 11 11

Non-Renewable Resources: Graphical Solution

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

40

36

32

28

24

20

16

12

8

4

Cumulative Resources Supplied

Cum

ulat

ive

Res

ourc

es

Weeks

Cumulative Resources Required

Resource Allocation Problem #3Issue: When is it better to “team” two or more

workers versus letting them work separately?

• Have 2 workers, Bob and Barb, and 4 tasks: A, B, C, D

• Bob and Barb can work as a team, or they can work separately

• When should workers be assigned to tasks? Which configuration do you prefer?

How to Assign Project Teams?

Configuration #1Bob and Barb work jointly on all four tasks; assume that they can complete each

task in one-half the time needed if either did the tasks individually

A C

B D

Start End

Configuration #2Bob and Barb work independently. Bob is assigned to tasks A and C; Barb is

assigned to tasks B and D

Bob and Barb: Configuration #1

TASK A TASK B TASK C TASK DDuration Prob Duration Prob Duration Prob Duration Prob

6 0.33 9 0.667 12 0.6 10 0.255 0.33 6 0.333 7 0.4 6 0.754 0.33

Expected duration 5.0 8.0 10.0 7.0

Configuration #1

Bob and Barb work jointly on all four tasks.

What is the expected project makespan?

Bob and Barb: Configuration #2Bob and Barb work independently. Bob is assigned to tasks A and C; Barb is

assigned to tasks B and D

Realization # A B C DBob

A + CBarb B + D

max (A+C, B+D) Prob

1 6 9 12 10 18 19 19 0.032 6 9 12 6 18 15 18 0.103 6 9 7 10 13 19 19 0.024 6 9 7 6 13 15 15 0.075 6 6 12 10 18 16 18 0.026 6 6 12 6 18 12 18 0.057 6 6 7 10 13 16 16 0.018 6 6 7 6 13 12 13 0.039 5 9 12 10 17 19 19 0.03

10 5 9 12 6 17 15 17 0.1011 5 9 7 10 12 19 19 0.0212 5 9 7 6 12 15 15 0.0713 5 6 12 10 17 16 17 0.0214 5 6 12 6 17 12 17 0.0515 5 6 7 10 12 16 16 0.0116 5 6 7 6 12 12 12 0.0317 4 9 12 10 16 19 19 0.0318 4 9 12 6 16 15 16 0.1019 4 9 7 10 11 19 19 0.0220 4 9 7 6 11 15 15 0.0721 4 6 12 10 16 16 16 0.0222 4 6 12 6 16 12 16 0.0523 4 6 7 10 11 16 16 0.0124 4 6 7 6 11 12 12 0.03

Bob and Barb: Configuration #2

Bob and Barb work independently. Bob is assigned to tasks A and C; Barb is assigned to tasks B and D

max (A+C, B+D) Prob

Cumulative Prob

12 0.07 0.0713 0.03 0.1015 0.20 0.3016 0.20 0.5017 0.17 0.6718 0.17 0.8319 0.17 1.00

Expected Project Makespan: 16.42

Parallel Tasks with Random Durations

START END

Task B

Task A

• Assume that both Tasks A and B have possible durations:

8 days with probability = 0.5

10 days with probability = 0.5

• What is expected duration of project? (Is it 9 days?)

Project Monitoring and Controln “It is of the highest importance in

the art of detection to be able to recognize, out of a number of acts, which are incidental and which are vital. Otherwise your energy and attention must be dissipated instead of being concentrated.”

Sherlock Holmes

Status Reporting?

One day my Boss asked me to submit a status report to him concerning a project I was working on. I asked him if tomorrow would be soon enough. He said, "If I wanted it tomorrow, I would have waited until tomorrow to ask for it!"

New business manager, Hallmark Greeting Cards

Control System Issues

n What are appropriate performance metrics?n What data should be used to estimate the value of each

performance metric?n How should data be collected? From which sources? At

what frequency?n How should data be analyzed to detect current and future

deviations?n How should results of the analysis be reported? To

whom? How often?

Controlling Project Risks

Key issues to control risk during projecct:

(1) what is optimal review frequency, and

(2) what are appropriate review acceptance levels at each stage?

“Both over-managed and under-managed development processes result in lengthy design lead time and high development costs.”

Ahmadi & Wang. “Managing Development Risk in Product Design Processes”, 1999

Project Control & System Variation

Common cause variation: “in-control” or normal variation

Special cause variation: variation caused by forces that are outside of the system

According to Deming:

• Treating common cause variation as if it were special cause variation

is called “tampering”

• Tampering always degrades the performance of a system

Control System Example #1n Project plan: We estimate that a task will

take 4 weeks and require n 1600 worker-hours

At the end of Week 1, 420 worker-hours have been used

Is the task “out of control”?

Control System Example (cont’d)Week 2: Task expenses = 460 worker-hours


370

380

390

400

410

420

430

440

450

460

470

1 2 3 4

Week

Cos

t (in

wor

ker-

hour

s)Week

Planned Cost (BCWS) Actual Cost

Cumulative Actual Cost

(ACWP)1 400 420 4202 400 460 880

Control System Example (cont’d)Week 3: Task expenses = 500 worker-hrs


WeekPlanned cost

(worker-hours)Actual cost

(worker-hours)Cumulative cost (worker-hours)

1 400 420 4202 400 460 8803 400 500 1380

0

100

200

300

400

500

600

1 2 3 4Week

Wor

ker-

hour

s

Earned Value Analysis

• Integrates cost, schedule, and work performed

• Based on three metrics that are used as the basic building blocks:

BCWS: Budgeted cost of work scheduled

ACWP: Actual cost of work performed

BCWP: Budgeted cost of work performed

Schedule Variance (SV)

Schedule Variance (SV) = difference between value of work completed and value of scheduled work

Schedule Variance (SV) = Earned Value - Planned Value

= BCWP - BCWS

Cost Variance (CV)

Cost Variance (CV) = difference between value of work completed and actual expenditures

Cost Variance (CV) = Earned Value - Actual Cost

= BCWP - ACWP

Earned Values Metrics IllustratedW

orke

r-H

ours

Week 1 Week 2 Week 3 Week 4 Week 5 Week 6

Present timeBAC

Actual Cost (ACWP)

Earned Value (BCWP)

Planned Value (BCWS)

Schedule Variance (SV)

Cost Variance (CV)

Relative Measure: Schedule Index

Schedule Index (SI ) = BCWPBCWS

If SI = 1, then task is on schedule

If SI > 1, then task is ahead of schedule

If SI < 1, then task is behind schedule

Relative Measure: Cost Index

Cost Index (CI) = BCWPACWP

If CI = 1, then work completed equals payments (actual expenditures)

If CI > 1, then work completed is ahead of payments

If CI < 1, then work completed is behind payments (cost overrun)

Example #2W E E K

1 2 3 4 5 6 7 8 9 10

6 6 6 8 10

12 12 12

10 10 12 12 12

Weekly Scheduled

Worker-Hrs 6 6 6 20 22 22 10 12 12 12Cumulative Scheduled

Worker-Hrs (BCWS) 6 12 18 38 60 82 92 104 116 128

Task A (36 worker-hrs)

Task B (36 worker-hrs)

Task C (56 worker-hrs)


Week 1 2 3 4 5Task A 15% 30% 40% 60% 80%Task B 25% 65%Task C Not started yet

Progress report at the end of week #5:

Cumulative Percent of Work Completed:

Worker-Hours Charged to Project:

Week 1 2 3 4 5Task A 5 6 8 10 10Task B 15 10Task C Not started yet

Example #2 (cont’d)Progress report at the end of week #5:

W E E K1 2 3 4 5 6 7 8 9 10

Cumulative Scheduled

Worker-Hrs (BCWS) 6 12 18 38 60 82 92 104 116 128

Actual Worker-Hrs Used (ACWP) 5 11 19 44 64

Earned Value (BCWP) 5.4 10.8 14.4 30.6 52.2

Schedule Variance (SV) -0.6 -1.2 -3.6 -7.4 -7.8Cost Variance

(CV) 0.4 -0.2 -4.6 -13.4 -11.8


0

20

40

60

80

100

120

140

1 2 3 4 5 6 7 8 9 10

Week

Perf

orm

ance

Met

ric

ACWP

BCWP

BCWS

Schedule Variance

Cost Variance

BAC

Using a Fixed 20/80 RuleCumulative Percent of Work Completed:

W E E K1 2 3 4 5 6 7 8 9 10

Cumulative Scheduled

Worker-Hrs (BCWS) 6 12 18 38 60 82 92 104 116 128

Actual Worker-Hrs Used (ACWP) 5 11 19 44 64

Earned Value (BCWP) 7.2 7.2 7.2 14.4 14.4Schedule

Variance (SV) 1.2 -4.8 -10.8 -23.6 -45.6Cost Variance

(CV) 2.2 -3.8 -11.8 -29.6 -49.6

Week 1 2 3 4 5Task A 20% 20% 20% 20% 20%Task B 20% 20%Task C Not started yet

Using a Fixed 20/80 Rule

0

20

40

60

80

100

120

140

1 2 3 4 5 6 7 8 9 10Week

Cos

t (in

Wor

ker-

hour

s)

BCWP

ACWPBCWS

Updating Forecasts: Pessimistic Viewpoint

= (64/52.2) 128 = 1.23 x 128 = 156.94 worker-hrs

Estimate at Completion (EAC) = ACWPBCWP

BAC = 1CI

BAC .

Assumes that rate of cost overrun will continue for life of project….

Updating Forecasts: Optimistic Viewpoint

Estimate at Completion (EAC) = BAC - CV = 128 + 11.8 = 139.8 worker -hrs .

Assumes that cost overrun experienced to date will cease and no further cost overruns will be

experienced for remainder of project life…

Multi-tasking with Multiple Projects

Project A Project B

A-1 B-1 A-2 B-2 A-3 A-4B-3 B-4

Consider two projects with and without multi-tasking

How to prioritize your work when you have multipleprojects and goals?

Due-Date Assignment with Dynamic Multiple Projects

• Projects arrive dynamically (common situation for both

manufacturing and service organizations)

• How to set completion (promise) date for new projects?

• Firms may have complete control over due-dates or only partial

control (i.e., some due dates are set by external sources)

• How to allocate resources among competing projects and tasks (so

that due dates can be realized)?

• What are appropriate metrics for evaluating various rules?

What Does the Research Tell Us?

• Study by Dumond and Mabert* investigated four due date assignment rules and five scheduling heuristics

• Simulated 250 projects that randomly arrive over 2000 days• average interarrival time = 8 days• 6 - 49 tasks per project (average = 24); 1 - 3 resource types• average critical path = 31.4 days (range from 8 to 78 days)

• Performance criteria: 1) mean completion time2) mean project lateness3) standard deviation of lateness4) total tardiness of all projects

• Partial and complete control on setting due dates* Dumond, J. and V. Mabert. “Evaluating Project Scheduling and Due Date Assignment Procedures: An Experimental Analysis” Management Science, Vol 34, No 1 (1988), pp 101-118.

Experimental Results

• No one scheduling heuristic performs best across all due date setting combinations

• Mean completion times for all scheduling and due date rules not significantly different

• FCFS scheduling rules increase total tardiness• SPT-related rules do not work well in PM (SASP)• Best to use more detailed information to establish due dates

Project Management Maturity Models

• Methodologies to assess your organization’s current level of

PM capabilities

• Based on extensive empirical research that defines “best

practice” database as well as plan for improving PM process

• Process of improvement describes the PM process from

“ineffective” to “optimized”

• Also known as “Capability Maturity Models”

PM Maturity Model Example*1) Ad-Hoc The project management process is described as disorganized, and occasionally

even chaotic. Systems and processes are not defined. Project success depends on individual effort. Chronic cost and schedule problems.

2) Abbreviated: Some project management processes are established to track cost, schedule, and performance. Underlying disciplines, however, are not well understood or consistently followed. Project success is largely unpredictable and cost and schedule problems are the norm.

3) Organized: Project management processes and systems are documented, standardized, and integrated into an end-to-end process for the company. Project success is more predictable. Cost and schedule performance is improved.

4) Managed: Detailed measures of the effectiveness of project management are collected and used by management. The process is understood and controlled. Project success is more uniform. Cost and schedule performance conforms to plan.

5) Adaptive: Continuous improvement of the project management process is enabled by feedback from the process and from piloting innovative ideas and technologies. Project success is the norm. Cost and schedule performance is continuously improving.

* source: The Project Management Institute PM Network (July, 1997), Micro Frame Technologies, Inc. and Project Management Technologies, Inc. (http://pm32.hypermart.net/)

Administración de Proyectos

Documents

Transcript of Administración de Proyectos