The Radar Cost Model

1
The Radar Cost Model

• Presented To
• The Society of Cost Estimating Analysis
• 2004 National Conference
• 15-18 June 2004
• Manhattan Beach, California
• Presented by
• Peter Meisl Kyle Ratliff
• MCR Federal, LLC

2
Overview

• Background
• Model Development
• Model Methodology
• Risk Assessment
• Key Cost Drivers
• User Input Sheet
• LCC Breakdown
• Radar Cost Model Uses
• Future Plans
• Summary
• Published Sources

3
Background

• Radar Cost Model
• Parametric cost engineering model
• Estimates the contractor and government Life
  Cycle Cost (LCC) of ground, sea, and airborne
  radars
• Evolved from the need to estimate a variety of
  radar programs within the Missile Defense Agency
  (MDA) and Marine Corps Systems Command (MCSC)
• Consists of Cost Estimating Relationships (CERs),
  analogies, and cost-to-cost factors tailored to a
  generic radar system Cost Element Structure (CES)
• Model ensures three critical user requirements
  are met
• Current program compatibility
• Flexible cost breakout
• Emphasis on technical detail

4
Model Hierarchy
Other Hardware
5
Model Development Approach

• Developed a generic CES
• Consistent with OSD MIL-STD Handbook 881B
• Populated CES with estimating methodologies from
  other sources (I.e., models, studies, planning
  factors, etc.)
• CERs
• Analogies
• Cost-to-Cost Factors
• Captured all input variables
• Identified 33 key parameters
• 14 are technical characteristics of radar
• Selected primary and alternate methodologies
• Developed new CERs as needed
• Populated model with THAAD-like parameters
• Successfully tested output against THAAD CDSRs

6
Model Flow

• Production CERs and Analogies yield Production T1
• Learning Curves used for projecting Recurring
  Manufacturing
• Below the Line (BTL) costs based on cost-to-cost
  factors
• Manpower estimate drives government costs
• Step Factor yields Prototype T1 (Dev T1)
• DevEng CER calculates Developmental Engineering
  cost based on
• Burn Rate (function of Dev T1)
• Dev Time (function of Dev T1 or user-entered)
• Software Development based on MCR CER
• OS costs
• A function of recurring manufacturing costs
  (Maintenance)
• A manpower build up for military personnel
• Schedule and Technical (S/T) scoring used to
  calculate risk by phase

7
Methodology Selection

• Four main methods
• CERs selected based on a set of criteria
  described below
• Most hardware other than electrical components
• Analogies to similar programs (SBX, GBRP, THAAD)
• Mostly for electrical components
• In-house models (e.g. T/R modules)
• Fit equations from other models (e.g. software
  based on SEER output)
• Use of cost-to-cost factors for BTL items
  requires source consistency
• Different sources use different WBSs,
  normalization techniques
• Necessary to ensure that percentages sum properly
• CER selection based on the following set of
  criteria
• Makes sense (signs correct)
• Good statistical fit
• Uses commonly known input variables
• Contains all (and only) logical cost drivers
• Distinguishes between types of radar, if
  appropriate
• Based on data taken from relevant systems

8
Risk Assessment

• Risk evaluated by scoring in key areas against a
  Schedule and Technical Risk Matrix
• Hardware
• Software
• IAT
• Mapping equations then translate scores into a
  distribution of expected outcomes through Monte
  Carlo simulation.

9
Hardware Risk Scoring Matrix
10
Mapping Risk Scores

• Once we have risk scores we must translate them
  into cost growth
• We achieve this through a set of mapping
  equations
• The dependent variable in these equations is the
  mean of the symmetrical triangular S/T
  distribution
• The low and high endpoints are derived from the
  variance of a regression on historical programs

11
Key Cost Drivers

• Number of Transmit/Receive (T/R) modules
• Peak power per T/R module
• Aperture
• Development time
• Duty cycle
• SLOC and E SLOC
• Radar quantity
• Military Personnel
• Contractor Support

12
User Input Sheet
13
LCC Breakdown
14
Radar Cost Model UsesEstimator or Cross-Check

• Radar-specific inputs (based on the CARD or
  analyst knowledge) can be entered into the model
• Perform drills for various radar excursions
• S/N/Poet Group
• STSS/SASC Radar/Sensor Study
• National Team Excursions
• Sensitivity analyses for changing antenna
  gain/power/aperture
• Cost impact of tradeoffs suggested by Radar Range
  Equation
• SLOC growth and New vs Used CSCI percentages
• Values may also be compared to Program Office or
  CCM estimates for total cost, or by line item

15
Radar Cost Model UsesCER Repository

• Model contains hundreds of CERs
• Up to 29 CERs per line item
• Analysts can discretely select CERs and use them
  in (independent) cost models

16
CER Data Sources

• MCR CER Database
• MCR Unified Missile/Radar CCDR Database (MS
  Access)
• MEADS Cost Sensor Briefing (Tecolote, 2002)
• Raytheon Company CPR DAS660-92-C-0184, DD for
  GBR-P, 3/25/99
• Raytheon Company CPR DASG60-92-C-0101 for THAAD,
  9/96
• THAAD and XBR CCMs
• Program Level Cost Factors for Radar Production
  Systems
• Vendor Quotes

17
Future Plans

• Main objectives
• Populate and test model with parameters of
  various radars
• Keep several CERs, and eliminate baseless ones,
  for each CES item
• Improve user interface to allow for CER selection
• Other possibilities
• Create spinoff models based on different types
  of radars (e.g. fixed site, ship-based)
• Develop CES correlation matrix to further reduce
  number of required input variables (if necessary)
• Use radar model as a template for future cost
  models (e.g. Missile Cost Model)
• Improve on weak areas of model
• Government costs
• OS costs
• Time phasing
• Migrate from MS Excel to ACEIT

18
Summary

• Effective parametric LCCM for estimating a
  variety of radar systems
• Consists of a repository of CERs, analogies, and
  Cost-to-Cost factors
• Simple interface based on user inputs
• Includes top and lower level outputs
• Flexible for running what-if drills, baseline
  excursions, ROMs, and LCCEs
• Capable of supporting DTC and CAIV efforts
• Successfully used on MDA and MCSC programs
• Evolving tool with enhancements in progress

19
Published Sources

• Boehm, Barry. Software Engineering Economics.
  (1981)
• Frederick, Brad C. and Vokokek, Alan S. CR-0074,
  Updated Cost Estimating Relationships For ABM
  Radars. (Tecolote, 1985)
• Cost Methodology Improvement for System Hazardous
  Waste/Disposal Costs. Contract Number
  SDIO84-93-C-0026, Task Order 001, Sequence No.
  A049, CDRL A004. (SAIC, 1994)
• Program Level Cost Factors for Radar Programs in
  Production. (MCR, 1990)
• Radar CERs Development and Production.
• Radar Hardware Cost Estimating Relationships
  Database. Seq A097, CRDL A005. (SAIC, 1995)
• Radar (Surface Based) Development Engineering
  Cost Estimating Relationship. (MCR, 1994)
• Recurring Production Support Factors. (Applied
  Research, 1993)
• Reinsenleiter, Vern. TN 00-02, Estimating
  Relationships for Development Engineering and
  Development Time. (MCR, 1999)
• TN-94-016. Updated RD Producibility Engineering
  Planning (PEP) CERs for Missile, Radar, and
  Other Types of Programs. (MCR, 1994)
• TN 98-05, BMD Radar Systems Cost Estimating
  Relationships. (MCR, 1998)
• TR-8740-2, Electro-Optical, Missile, Radar and
  Avionics System Cost Research Cost Analysis
  Techniques Report Volume 1 Radar Production
  Cost Model. (MCR, 1988)
• TR 9001-1, Development Engineering
  Below-The-Line Development Cost Models. (MCR,
  1990)
• Yates, Edward H. Waller, W. Eugene and Vaughan,
  Lern G. A Parametric Approach to Estimating the
  Cost of Development Engineering. ARI/87 TM-387,
  Contract No. DASG 60-84-C-0061. (Applied
  Research, 1987)

20
Back-Up Charts
21
The Mapping Equation

• The Mapping Study produced separate equations for
  both RDTE and Procurement

X Risk Score