1. Modeling the Effect of Budget Constraints
on Cost and Schedule
NASA 2012 PM Challenge
Darren Elliott – Tecolote Research, Inc.
22 February 2012
Los Angeles Washington, D.C. Boston Chantilly Huntsville Dayton Santa Barbara
Albuquerque Colorado Springs Goddard Space Flight Center Johnson Space Center Ogden Patuxent River Washington Navy Yard
Ft. Meade Ft. Monmouth Dahlgren Quantico Cleveland Montgomery Silver Spring San Diego Tampa Tacoma
Aberdeen Oklahoma City Eglin AFB San Antonio New Orleans Denver Vandenberg AFB
PRT#115 22 February 2012 Approved for Public Release
3. Evolution of Cost/Schedule
Understanding
Implementation of cost risk analysis to provide a better gauge of the most likely
required effort for a project
NASA established policy to fund at a target cost confidence level
Reality was that cost risk funds exceeded targeted budget levels
Led to additional analysis on best techniques for applying reserve and setting up the overall
project budget
Recognition that enhanced modeling was needed, as many costs are time-
dependent (e.g., fixed infrastructure, program support, systems engineering) in
nature
Developed methods to address costs based on time behavior (e.g., time-dependent and time-
independent)
Enhanced understanding hat additional costs are incurred due to schedule delays
associated with alignment of work packages
Developed methodology to integrate cost and schedule risk analysis to determine the joint
confidence level (percent chance of meeting both cost and schedule objectives)
NASA established policy to fund projects at a target joint confidence level
Realization that funding is a major driver to schedule and total cost
Lack of budget availability stretches schedule
Currently researching and developing methods to address this problem
PRT#115 22 February 2012 Approved for Public Release 3
4. Our Reality is an Integrated System;
Where Budget Availability is a Major Input
PRT#115 22 February 2012 Approved for Public Release
5. Several Tools Have Recently Been Developed
to Assess Impact of Funding Constraints
Tool 1: Deterministic Based Approach
Sponsored by NASA JSC , developed by Tecolote Research, Inc
Requires identification of major cost elements, breakdown into TD/TI behavior, and notional
understanding of dependency (serial or parallel) between cost elements
Uses time-phased cost and budget information
Runs in Excel
Tool 2: Cost Risk Based Approach
Sponsored by NASA HQ, developed by Tecolote Research, Inc
Requires development of a cost-risk analysis
Uses cost risk statistics, and time-phased cost and budget information
Runs in ACE or Excel
Tool 3: Integrated Cost and Schedule Risk Analysis Approach
Sponsored by NASA JSC and NASA HQ , developed by Tecolote Research, Inc
Requires development of an integrated cost and schedule risk analysis, allocation of costs to
schedule effort, and breakdown of cost into TD/TI behavior
Uses schedule logic, cost allocation to schedule activities, time-phasing of costs, cost and/or
schedule statistics, allocation of schedule/cost effort into budget items, time-phased budget
information, and can incorporate discrete threats
Runs in Excel /Crystal Ball
PRT#115 22 February 2012 Approved for Public Release 5
6. Modeling the Effect of Budget Constraints on Cost and Schedule
DETERMINISTIC ANALYSIS
PRT#115 22 February 2012 Approved for Public Release 6
7. Deterministic Analysis Model
Purpose: Provide program managers and resource
analysts the ability to assess the effect that budget
constraints will have on a program
Requires a simple set of data
Includes estimated effect on both cost and schedule
Allows for rapid simulation and comparison of different
scenarios
Requires cost plan for major activities/efforts and a
notional understanding of major effort dependencies (e.g.,
serial or parallel activities)
PRT#115 22 February 2012 Approved for Public Release
8. Model Overview
The model attaches Time-Independent (TI) and Time-
Dependent (TD) costs to schedule elements
TI cost elements: Total cost is constant regardless of duration
TD cost elements: Total cost increases as duration increases (labor
rate and resources)
Budget constraints reduce spending and cause the
associated schedule elements to extend
Assumes a constant spend rate within each fiscal year
The spend rate is determined from the year’s constraint
TD spend rates are considered fixed, while TI spend rates are
allowed to vary to fit within the constraint
PRT#115 22 February 2012 Approved for Public Release
9. Model Overview
Named
Schedule Cost Budget Excursion
Elements Elements Constraint Budget Results
s Constraint
Engine
Results
Display
Engine
Schedul
e
Element
Referenc
e
Analyze Show
User Interface
PRT#115 22 February 2012 Approved for Public Release
10. Example Analysis: Baseline Case
Four budget elements, includes series and parallel events
Summary Results Gantt Chart
Date
10/1/2012 10/1/2013 10/1/2014 10/1/2015 9/30/2016
Task 1
Task 2
Start Date Finish Date
Task 3
Task 1 10/1/2012 2/13/2016
Task 2 4/1/2013 1/25/2014
Task 4
Task 3 1/26/2014 6/18/2015
Task 4 1/26/2014 11/25/2015
• Each task has an associated TI and TD cost, resulting in the
following spending profile: Summary Results by Schedule Elements (Constant Year $)
250
This example assumed 200
up to $200 was
available per year, as 150
Task 4
demonstrated by the
$M
Task 3
Task 2
black line 100
Task 1
50
0
2011 2012 2013 2014 2015 2016 2017 2018
PRT#115 22 February 2012 Approved for Public Release
11. Example Analysis: Excursion 1
• What if the budgets are constrained by 10% each year?
– Because tasks will now begin slipping, it is also necessary to define constraints
in the out-years beyond 2016, this example assumes funding continues at 2016
levels ($180 in this example)
Summary Results Gantt Chart Summary Results by Schedule Elements (Then Year $)
200
Date
10/1/2012 10/1/2013 10/1/2014 10/1/2015 9/30/2016 9/30/2017 180
160
Task 1 140
120
Task 4
$M
100 Task 3
Task 2
80 Task 2
Task 1
60
Task 3
40
20
Task 4 0
2011 2012 2013 2014 2015 2016 2017 2018 2019
Summary Results Gantt Chart - Original Budget
Date
10/1/2012 10/1/2013 10/1/2014 10/1/2015 9/30/2016 9/30/2017
Original Duration Excursion Duration Original Cost Excursion Cost
Task 1
Task 1 1230 1541 $402.64 $453.18
311 day slip from Task 2 299 356 $69.35 $72.70
Task 2 baseline
Task 3 508 666 $57.99 $63.28
Task 4 668 874 $144.35 $170.78
Task 3
Task 4
PRT#115 22 February 2012 Approved for Public Release
12. Example Analysis: Comparison
Excursion has a 12% increase in cost
All excursion tasks have an increased duration
Task 1 has the largest increase in duration
Increases in duration correspond to increased TD costs
Constraining the budget will avoid cost in those years, but the
deferred work and presence of fixed costs will result in a net cost
increase and schedule slip
How can this information be used? Some examples:
It is apparent Task 1 is a large cost driver in the new case, prioritizing it at the
expense of the other, shorter, tasks may result in a total cost savings
If the goal is a budget reduction, it may be preferable to focus the reduction on only
one or two years instead of spreading it evenly
These different scenarios can be run to explore the nature of the
budget constraint
PRT#115 22 February 2012 Approved for Public Release
13. Example Analysis: Excursion 2
Prioritizing Task 1 allows it to finish earlier, and moves the
entire project back to the left
Summary Results Gantt Chart
Date
10/1/2012 10/1/2013 10/1/2014 10/1/2015 9/30/2016 9/30/2017
Task 1 Excursion 1 Finish Date
Task 2
Baseline Finish Date
Task 3
• Task 1’s earlier
Task 4
completion frees up
money for Tasks 3 and 4
• Final cost in this case: $708 Summary Results by Schedule Elements (Constant Year $)
– Final cost in Excursion 1: $760 160
– Final cost in Baseline: $674 140
120
100
Task 1
$M
80 Task 2
Task 3
60
Task 4
This is an example of how the tool can be 40
used to potentially optimize cost and 20
0
schedule within a budget constraint 2011 2012 2013 2014 2015 2016 2017 2018
PRT#115 22 February 2012 Approved for Public Release
14. Example Analysis: Excursion 3
What if some of the budget cuts can be reallocated?
In this example, we will cause FY2013 to take most of the budget cuts, but
then increase the budget in the out-years:
2013 2014 2015 2016 2017
$100.00 $200.00 $200.00 $200.00 $200.00
Summary Results Gantt Chart
Date
10/1/2012 10/1/2013 10/1/2014 10/1/2015 9/30/2016 9/30/2017
Task 1
Task 2
Task 3
Task 4
• This example finishes at nearly the same time as the original excursion
– Total cost reduced to $743, compared to the excursion 1’s $760
PRT#115 22 February 2012 Approved for Public Release
15. Modeling the Effect of Budget Constraints on Cost and Schedule
COST RISK ANALYSIS
PRT#115 22 February 2012 Approved for Public Release 15
16. Current Techniques to Align Cost and Schedule
via Uncertainty Analysis
Total - Cost Estimate Uncertainty Analysis Results BY2010 $M
100.0% 4.5%
90.0% 4.0%
80.0% 3.5%
Probability (Histogram)
70.0%
Confidence Level (CDF)
3.0%
60.0%
2.5%
50.0%
2.0%
40.0%
1.5%
30.0%
20.0% Time Phased - Risk Adjusted Estimate 1.0%
Technical $100
10.0% 0.5%
Parameters $90
0.0%
$75 $215 $356 $497 $638 $778 $919 $1,060 $1,201 $1,341
0.0%
Allocated Dollars for 70% CL
$80 Probability Histogram Confidence Level (CDF)
CERs Point Estimate
$70 Cost Uncertainty Analysis
$60
TY $M
$50
$40
$30
$20
Project Finish Date
$10
100.0% 2.0%
$0
90.0% 1.8%
80.0%
2006 2007 2008 2009 2010
1.6%
2011 2012 2013 2014 2015 2016
70.0% 1.4%
Probability (Histogram)
Confidence Level (CDF)
60.0% 1.2%
50.0% 1.0%
40.0% 0.8%
30.0% 0.6%
20.0% 0.4%
10.0% 0.2%
0.0% 0.0%
24 Mar 2012 24 Dec 2012 24 Sep 2013 24 Jun 2014 24 Mar 2015 24 Dec 2015
Probability Histogram Confidence Level (CDF)
Schedule Schedule Uncertainty Analysis
PRT#115 22 February 2012 Approved for Public Release
17. However, Budget Profiles Rarely Match
Risk-Adjusted Time-Phased Estimates
$100
$90
Allocated Dollars for 70% CL
$80
Shortfall Point Estimate
$70
Project Budget
$60
TY $M
$50
$40
$30
$20
$10
$0
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Annual budget sufficient to cover estimated point estimate effort
Shortfall in funding 70% effort for years 2007-2011
How Does Shortfall Impact Project?
PRT#115 22 February 2012 Approved for Public Release
18. The Concept
Tecolote Developed a ROM-level analysis technique for NASA to
gauge the impact of budget availability on a project’s target cost
confidence level
The Technique requires:
Risk adjusted, time-phased cost estimate
Annual budget information
User input on how to address multiple items (e.g., penalties, etc)
The General Approach
Compare estimated effort (i.e., point estimate, risk adjusted time
phased results, or annual risk iteration results) to available budget
Identify and track budget shortfalls
Rollover unfunded effort, with associated inflation and productivity
penalties, to future years
Apply logic to use available budget to fund rollover effort
PRT#115 22 February 2012 Approved for Public Release
19. General Approach for Three
Different Scenarios
PRT#115 22 February 2012 Approved for Public Release
20. User Inputs and Controls
The user has the ability to:
Select type of analysis to
conduct
Point estimate
Risk-adjusted (e.g., 70%) cost
estimate
Dynamic assessment of
confidence level results
Specify budget scenarios Advanced Considerations
Extend budget at peak
Incorporating fixed costs (LOE) into
Infuse/Reduce funds in specific
consideration
year
Conducting portfolio analysis
Allow budget carryover
Incorporate penalties for rollover Outputs
effort: Initial phasing result for Target CL
Inflation considerations Constrained phasing result
Productivity loss Additional years of funding required
PRT#115 22 February 2012 Approved for Public Release
21. Example Case – 70% Effort
Exceeds Available Budget
Phased budget and point estimate
TY$M FY 2007 FY 2008 FY 2009 FY 2010 FY 2011 FY 2012 FY 2013 FY 2014 Total
Budget $25 $55 $65 $80 $70 $60 $30 $15 $400
Cost $21.8 $51.9 $63.6 $62.4 $52.9 $37.8 $18.9 $1.8 $311
Cost Estimate Uncertainty Analysis Results TY $M
100.0% 4.5%
Cost risk analysis data (TY$M) 90.0%
80.0%
4.0%
3.5%
Probability (Histogram)
Point Confidence Standard 70.0%
Confidence Level (CDF)
3.0%
Mean CV 60.0%
Estimate Level Deviation 50.0%
2.5%
2.0%
40.0%
$311 42% $372 $168 0.45 30.0%
1.5%
20.0% 1.0%
10.0% 0.5%
0.0% 0.0%
$75 $216 $357 $499 $640 $781 $922 $1,063 $1,205 $1,346
Probability Histogram Confidence Level (CDF)
Cost estimate @ 70% confidence level
TY$M FY 2007 FY 2008 FY 2009 FY 2010 FY 2011 FY 2012 FY 2013 FY 2014 Total
70% CLE $29.8 $71.2 $87.1 $85.6 $72.4 $51.7 $25.9 $2.5 $426
Budget Shortfall to Fund 70% CLE
PRT#115 22 February 2012 Approved for Public Release
22. Understanding the Shortfall – Work
Slips to the Right
Budget exceeds
Budget vs 70% Risk-Adjusted Estimate - Surplus/Shortfall
Budget vs Pt Estimate - Surplus/Shortfall Total Budget
phased point
$25 estimate in every year
inadequate to
$26M Total Shortfall -
$89M Budget Surplus Limited70% CLE
fund surplus in
$20 early years
Funds Available For
Reserve Utilization
Largeglance seems
At first
shortfall
$15
in early years, if
that enough reserves
Annual Surplus / Shortfall - TY$M
are available for
$10
funds
program cannot
be obtained,
$5
$5 effort will slip
into future
$0
$0
2007
2007 2008
2008 2009
2009 2010
2010 2011
2011 2012
2012 2013
2013 2014
2014
periods
-$5
-$5
2007 work
slips to 2008
-$10 and so on…
$50M of work effort Extended work
-$15
cannot be done in carries
2007-2011 penalties
-$20
Inflation
-$25 Productivity
What are Possible Budget Scenarios?
PRT#115 22 February 2012 Approved for Public Release
23. Identifying Budget Scenarios
Budget Scenario Considerations
Need to be realistic
Near-term funds are difficult to obtain
Annual increase must match capability to ramp up staffing levels and should track
to required work
Should not have extreme changes year-to-year
Difficult to increase beyond peak spending year
Cannot upset overall portfolio needs
Potential Options
Identify infusion of funds into specific years
Extend funding beyond peak funding year at or near peak value
$100.000
$90.000
$80.000
$70.000 Project Budget
$60.000
$50.000
$40.000
Budget Extension
at Peak
$30.000
$20.000
$10.000
$0.000
2007 2008 2009 2010 2011 2012 2013 2014 2015
PRT#115 22 February 2012 Approved for Public Release
24. Example Case Results – Effort Rollover
Fits Under Budget Constraint
$100
70% Risk Adjusted Estimate
RolloverEffort
70% Risk Adjusted Estimate
$90 14
2
3 Budget Funded Effort
BudgetEstimateEffort
Point Funded Effort
Budget Funded
$80
70% Risk Adjusted Estimate
Project Budget
Extended Budget
$70 Extended Budget
$60
$50
$40
$30
$20
$10
$0
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
1. Initial Conditions indicate budget inadequate to fund 70% CLE
2. Project funded to budget value
3. Budget scenario created to extend budget at peak value and 70%
time-phased estimate funded to extended budget values
4. Impact of rollover effort funded in out-years
PRT#115 22 February 2012 Approved for Public Release
25. Modeling the Effect of Budget Constraints on Cost and Schedule
INTEGRATED COST AND
SCHEDULE RISK ANALYSIS
PRT#115 22 February 2012 Approved for Public Release 25
26. Integrated Cost/Schedule Budget
Constrained Reserve Phasing Model
A tool is needed that accurately models the relationships
between the work to be done, the annual budget available,
and the overall cost and finish date of a project
Tool can be used to determine appropriate annual
budgets and reserve strategies to meet the Joint
Confidence Level requirement
The tool is called Budget-Constrained Reserve Phase
(BCRPhase)
26
PRT#115 22 February 2012 Approved for Public Release
27. High - Level Model Concept
Schedule Activities from JCL Models (or Project Schedules) are
Mapped into Budget Organizations (Groupings)
Links Between Budget Organizations are Identified
Costs from JCL Models (or Projects Costs/Budget) are Mapped
into Behavior Buckets within Budget Organizations
Cost Distributions Obtained from JCL Models (or other Analytical
Methods) are Defined for Each Element
Discrete Risks from JCL Models (or Project Risk Lists) are
Specified
Annual Values are Identified for Each Budget Organization
Monte Carlo Simulations are Generated to Determine Effort, Work
is Adjusted to Fit Effort within Budget Level
PRT#115 22 February 2012 Approved for Public Release 27
28. Modeling Approach – Enabling
via Budget Organization Levels
PRT#115 22 February 2012 Approved for Public Release 28
29. Example Simple Project Model
(5 Budget Organization Items)
Budget
Effort/Costs
Schedule
PRT#115 22 February 2012 Approved for Public Release 29
30. Model in Action – What Happens When
Effort Exceeds Budget Availability
Organization 1
Organization 2
Time
Time-Driving Costs (Touch Labor, Materials, etc.) Schedule Activity
Time-Driving Costs Driven By Schedule Slip Milestone
Level Of Effort Costs Schedule Dependency
PRT#115 22 February 2012 Approved for Public Release 30
31. Modeling the Effect of Budget Constraints on Cost and Schedule
SUMMARY
PRT#115 22 February 2012 Approved for Public Release 31
32. Summary
Understanding connection between cost and schedule is of utmost
importance for estimators
Available funds and resources are major items that directly impact
overall project schedule and end costs
Tools have been developed and research is continuing to help us
assess this intricate system
Budget
Duration
Technical
Reqts Work Plan
Scope
Cost
Effort Size
PRT#115 22 February 2012 Approved for Public Release 32