Schedule Risk Analysis (SRA) Tariq Ashraf Systems Engineering Feb 24, 2005.

Schedule Risk Schedule Risk Analysis (SRA)Analysis (SRA)

Tariq AshrafTariq Ashraf

Systems EngineeringSystems Engineering

Feb 24, 2005Feb 24, 2005

Why SRA?Why SRA?

Prior to program execution, management needs to Prior to program execution, management needs to know what are the potential schedule road blocks know what are the potential schedule road blocks ahead of him/herahead of him/her– Establish warning indicators up frontEstablish warning indicators up front– Build mitigation into the IMS during planningBuild mitigation into the IMS during planning

Deterministic model fallacy Deterministic model fallacy – Critical path based upon linked activities with Critical path based upon linked activities with

the least amount of float the least amount of float – Assumes all linked activities are low risk – not Assumes all linked activities are low risk – not

true in real lifetrue in real life

ProgramAttention

ProgramAttention

ProgramAttention

Risk Impact on Critical PathRisk Impact on Critical Path

Activity X Potential Critical Path

Probabilistic Model

Activity 1 Activity 2 Activity 3

Activity 4 Activity 5 Activity 6

Activity 7 Activity 8 Activity 9

9-10-12 14-15-17 19-20-22 51

14-15-17 9-10-12 14-15-17 51

9-10-12 9-10-12 19-20-34 58

High RiskActivity

Deterministic Model

Activity 1 Activity 2 Activity 3

Activity 4 Activity 5 Activity 6

Activity 7 Activity 8 Activity 9

10 15 20 45

15 10 15 40

10 10 20 40

Activity X Critical Path

Brief ObjectiveBrief Objective

Describe a structured methodology for conducting Schedule Risk Analysis (SRA)

Schedule Risk Assessment Schedule Risk Assessment MethodologyMethodology

Two step process:Two step process:– Step 1 : Assess and determine the risk level of each Step 1 : Assess and determine the risk level of each

IMS activity to identify activities that are high or IMS activity to identify activities that are high or moderate/medium risk – moderate/medium risk – the most important element of the most important element of SRA SRA

– Step 2 : Determine potential impact of the Step 2 : Determine potential impact of the High/Moderate risk activities on program events using a High/Moderate risk activities on program events using a simulation tool simulation tool

Step 1. How to Determine the Step 1. How to Determine the Risk Level of an Activity?Risk Level of an Activity?

Assess each activity against the Assess each activity against the elements that impact the activityelements that impact the activity– The activity model provides the answerThe activity model provides the answer

Activity ModelActivity Model

ActivityActivityInputInput

Dependency

Input Dependency – One Input Dependency – One PredecessorPredecessor

Ps = 80%Activity 1(Ps = 80%)

Activity 2

Ps = Probability of Success

Input Dependency – Multiple Input Dependency – Multiple Concurrent PredecessorConcurrent Predecessor

The greater the number of concurrent predecessor activities,the lower the Cumulative Probability of Success

Activity 1(Ps = 80%)

Activity 2(Ps = 80%)

Activity 3(Ps = 80%)

Activity 4(Ps = 80%)

Activity 5

Cumulative Ps

80% x 80% x 80% x 80% = 41%

Ps = Probability of Success

Ps = 41%

Activity Model (continued)Activity Model (continued)

ActivityActivityInputInput

Dependency

Source

Timeliness of input is dependent uponcontrol exercised over the input source

Activity Model (continued)Activity Model (continued)

ActivityActivity

Res

ou

rce

Res

ou

rce

InputInput

Dependency

Source

Constraint

Resource ConstraintResource Constraint

There is a loss in resource efficiency when a resourceis applied to multiple concurrent activities

Activity 1

Resource A

Activity 2

Elapsed Time

Setup

Work on Task

Suspend

Non Productive Work

Productive Work

Theory of Constraint

Activity Model (continued)Activity Model (continued)

EnvironmentEnvironment

ActivityActivity

Res

ou

rce

Res

ou

rceP

rocess

Pro

cess

InputInput OutputOutput

Dependency

Source

Constraint

Maturity Complexity

Expectancy

Source

State

So Now What?So Now What?

Develop a likelihood rating matrix for each Develop a likelihood rating matrix for each activity element activity element – Example:Example:

Input DependencyInput Dependency– 1 – No Predecessor1 – No Predecessor– 5 – N Concurrent Predecessors5 – N Concurrent Predecessors

Resource StateResource State– 1 – Existing1 – Existing– 5 – New development using state-of-the art5 – New development using state-of-the art

Likelihood MatrixLikelihood Matrix

Dependency Source Maturity Complexity Constraint Source State

5>7 Concurrent Predecessor Activities

Outs ide Agency New ProcessExtremely Complex

Resource dedicated to more

than two concurrent activities

Outs ide Agency

New resource to be developed

using s tate-of-the art technology

Low Confidence that output will

meet expectation

Fully Dependent (100% of duration)

45-6 Concurrent

Predecessor Activities

CustomerMajor

modification to exis ting Process

Significantly Complex

Resource dedicated to more than two activities

Customer

New resource to be developed using exis ting technology

Insufficient data to make

judgement

Major Dependency (75%

of duration)

33-4 Concurrent

Predecessor Activities

Within Corporation

Moderate modification to exis ting process

Moderately Complex

Resource dedicated to two

concurrent activities

Within Corporation

Major Modification to

exis ting resource

Slight Confidence, based on past

experience, that output will meet

expectation

Moderately (50% of duration) Dependent

21-2 Concurrent

Predecessor Activities

Subcontractor OR Supplier

Minor modification to exis ting process

Minor ComplexityResource

dedicated to two activities

Subcontractor OR Supplier

Minor Modification to

the Resource

Moderate Confidence, with backup data, that out put will meet

expectation

Minor (25% of duration)

Dependency

1 No Predecessor Activity

Within Program Process Exis ts SimpleResource

dedicated to this activity

Within Program Resource Exis ts

High Confidence, with backup data, that out put will meet expectation

Not Dependent

Likelihood FactorsLevel Output

ExpectencyEnvironment Dependency

Input Process Resource

EXAMPLE

So Now What?So Now What?Now that the activity elements have been identified develop a likelihood rating Now that the activity elements have been identified develop a likelihood rating matrix for each activity element matrix for each activity element – Example:Example:

Input DependencyInput Dependency– 1 – No Predecessor1 – No Predecessor– 5 – N Concurrent Predecessors5 – N Concurrent Predecessors

Resource StateResource State– 1 – Existing1 – Existing– 5 – New development using state-of-the art5 – New development using state-of-the art

For each activityFor each activity– Using the likelihood rating matrix, determine the level for each activity element Using the likelihood rating matrix, determine the level for each activity element – Determine the schedule consequence using the criteria provided in the risk planDetermine the schedule consequence using the criteria provided in the risk plan– Determine the activity risk level by mapping highest likelihood level to schedule Determine the activity risk level by mapping highest likelihood level to schedule

consequenceconsequenceEstablish Max-Min durations criteria for high, moderate, low risk activityEstablish Max-Min durations criteria for high, moderate, low risk activity– DurationsDurations– ++ of % of ML of % of ML

The risk level determines the The risk level determines the estimate confidence and estimate confidence and

duration spread (Min-ML-Max) duration spread (Min-ML-Max) of an activity of an activity

High Risk Activity

Min Duration Spread

MaxMost Likely

Confidence Level Low High

Moderate/Medium Risk Activity

Min

Duration Spread

Max

Most Likely

Confidence Level Low High

Low Risk Activity

Low Confidence Level High

Most Likely

Duration Spread

MinMax

Step 2. Determining the Impact of High/Moderate Risk Step 2. Determining the Impact of High/Moderate Risk Activities on Program Event(s)Activities on Program Event(s)

Review the IMS logic for the following:Review the IMS logic for the following:– All activities are linkedAll activities are linked– The logic makes senseThe logic makes sense

Maximize FS relationships. Simulation models do not like other Maximize FS relationships. Simulation models do not like other relationships (SS, SF, FF). Also minimize lags in the activity logic i.e. relationships (SS, SF, FF). Also minimize lags in the activity logic i.e. FS+20, FS-20, etc.FS+20, FS-20, etc.For each activity, input estimates (Minimum , Most Likely, Maximum ) based For each activity, input estimates (Minimum , Most Likely, Maximum ) based upon the risk level of the activity and run simulationsupon the risk level of the activity and run simulationsOut put generated by simulation toolOut put generated by simulation tool– Probability vs. Time curve for each event selected for analysisProbability vs. Time curve for each event selected for analysis– Criticality Index for each task (during the simulation, the number of times an Criticality Index for each task (during the simulation, the number of times an

activity fell on the critical path)activity fell on the critical path)Results from the first simulation may not meet expectationResults from the first simulation may not meet expectationAdjust IMS and reiterateAdjust IMS and reiterate

ConclusionConclusionRisk is inherent in a complex system with multiple Risk is inherent in a complex system with multiple stakeholdersstakeholders

Prior to execution, management must know what Prior to execution, management must know what risks are embedded in the IMSrisks are embedded in the IMS

Applying a structured methodology ensures schedule Applying a structured methodology ensures schedule risk is identified consistently by all stakeholdersrisk is identified consistently by all stakeholders

Analysis is time consuming and requires input from Analysis is time consuming and requires input from all stakeholders but the value cannot be all stakeholders but the value cannot be underestimated underestimated