Lean Aerospace Initiative Plenary Workshop Models for Analyzing Complex System Development Cost, Schedule, and Performance Risks October 14, 1998 Presented By: Tyson R. Browning Product Development Focus Team Research Sponsored By Lean Aerospace Initiative Lean Aerospace Initiative l Who is This For? Decision makers needing to balance project cost and schedule and product technical performance l Project planning l Project management and control l Proposal preparation l Design reviews l Project replanning Tyson R. Browning — 2 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative l Motivation l Approach and goals l New results Overview – Model one: cost and schedule risk – Data: UCAV process – Model two: cost, schedule, and performance (CS&P) risk l Product development management insights Tyson R. Browning — 3 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative l l l Research Motivation Competitive advantage, Maximum return on investment, and Maximum product value …depend on balancing… l l l Product price (or life cycle operating cost) Delivery timing “Technical” performance “Overall Performance” or “Value to Customer” => Establish and maintain balance of product performance dimensions => Establish and maintain balance in development process => Understand sources and impacts of uncertainty (risk) Tyson R. Browning — 4 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative CS&P Trade-offs Cost Schedule Technical Performance l l l Different process configurations have different C&S outcomes They may also deliver differing levels of expected performance We need a way to help navigate these choices Tyson R. Browning — 5 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative Research Approach and Goals l Observe product development processes to determine sources of uncertainty and risk l Develop methodology and model for quantifying CS&P risks in PD processes l Use model to provide insights for improved PD process planning and management Tyson R. Browning — 6 Observation Modeling Improvement Understanding PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative Design Structure Matrix Preliminary Design Activities (Example Set) A4 A511 A512 A531 A521 A522 A5341 A532 A533 A5342 A5343 A5344 A54 A6 Prepare UCAV Preliminary DR&O Create UCAV Preliminary Design Configuration Prepare & Distribute Surfaced Models & Int. Arngmt. Drawings Perform Aerodynamics Analyses & Evaluation Create Initial Structural Geometry Prepare Structural Geometry & Notes for FEM Develop Structural Design Conditions Perform Weights & Inertias Analyses Perform S&C Analyses & Evaluation Develop Balanced Freebody Diagrams & Ext. Applied Loads Establish Internal Load Distributions Evaluate Structural Strength, Stiffness, & Life Preliminary Manufacturing Planning & Analyses Prepare UCAV Proposal • • • • • • • 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 2 3 4 5 6 7 8 ■ ■ ■ ■ ■ ■ ■ ■ 9 10 11 12 13 14 ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Square matrix showing relation of each activity to every other Row: information required (inputs) Column: information provided (outputs) Time-based ordering of activities Counter-clockwise flow of information between activities in process Alternating bands show concurrency Superdiagonal marks show iteration/rework Tyson R. Browning — 7 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative l l A Model to Quantify Cost and Schedule Risk DSM representation Activity-based – Stochastic durations and costs – Learning curve effect on rework l Accounts for rework caused by timing of information exchange – Probabilistic iteration – Variable impacts (% rework) l l l Dynamic, Monte Carlo simulation Activity concurrency (info.-based) Outputs: cost and schedule outcome distributions and risk factors Tyson R. Browning — 8 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative Model Algorithm (Simplified) Rework Probabilities 1 2 3 4 5 6 7 A4 A511 A512 A531 A521 A522 A5341 A532 A533 Prepare UCAV Preliminary DR&O 1 Create UCAV Preliminary Design Configuration 2 .4 Prepare & Distribute Surfaced Models & Internal Arrangement 3 .5 Drawings .4 Perform Aerodynamics Analyses & Evaluation 4 .3 .5.5 .3 .5 Create Initial Structural Geometry 5 .4 .5 .1 .4 Prepare Structural Geometry & Notes for FEM 6 .1 .4.8 .1 .1 .4 Develop Structural Design Conditions 7 .4 .1 Perform Weights & Inertias Analyses & Evaluation8 .5 Perform S&C Analyses & Evaluation 9 .4 .5.5 .5 A5342 A5343 A5344 A54 A6 Develop Balanced Freebody Diagrams & External10Applied Loads.1 .3 Establish Internal Load Distributions 11 Evaluate Structural Strength, Stiffness, & Life 12 .4 Preliminary Manufacturing Planning & Analyses 13 .5 Prepare UCAV Proposal 14 .3 .4 .4.5 .4 .3 .5 .3 .5 .1 .4 .8 9 10 12 13 14 W WN 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 .1 .1 .3 .1 .2 .3 .5 .5 .3 .4 .5 .4 .4 .8 .8 .4 .4 .9 .3 .4 .8 .8 .8 .5 .4 .5 .5 .1 .5 .8 .5 .5 .1 .4 .5 .1 .0 .3 .5 .5 .8 11 .2 .3 .5 .3.5 .9 .5 8 .3 .4 .3.4 .4 .5.4 .5.4 .3 .8 .8 .8 .8 .8 Rework Impacts l l l Based on time steps Sample activity durations (convert to time steps) and costs Each time step – Determine band of concurrent activities (WN) – Do work; add time and cost (S and C) – Check for iteration and second-order rework; add impact to W l Simulate hundreds of runs to generate C, S sample pairs Tyson R. Browning — 9 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative l l l l Uninhabited Combat Aerial Vehicle (UCAV) at Boeing Conceptual and preliminary design Collected through interviews and surveys Process data – – – – – – l l UCAV Process Report Activities Durations Required resources Learning curve effect on rework Inputs and outputs (information flow) Sensitivities to and impacts of input changes Serves as a process knowledge base (for organizational learning) Data represented using DSM Tyson R. Browning — 10 PD Browning 101498 ©1998 Massachusetts Institute of Technology Sample Result for Single Run of UCAV Preliminary Design Process Lean Aerospace Initiative 20 40 60 80 100 120 132 1 2 3 4 5 6 7 8 9 10 11 12 13 14 l l l l Visualize critical path Nominally sequential process Limited concurrency (dynamic) Waiting on results when appropriate Tyson R. Browning — 11 PD Browning 101498 ©1998 Massachusetts Institute of Technology Results for UCAV Preliminary Design Process Lean Aerospace Initiative Target 1.0 .8 PC(un.) ≈ 0.24 .6 .4 .2 .0 570 600 630 660 690 720 750 780 810 840 Cost ($k) Target 1.0 0.8 0.6 0.4 • Distributions (not point values) and targets —> P(unacceptable) • with Impacts (e.g., quadratic loss function) —> RC & RS 0.2 0.0 120 126 132 138 144 150 156 162 168 174 180 PS(un.) ≈ 0.91 Schedule (days) Tyson R. Browning — 12 PD Browning 101498 ©1998 Massachusetts Institute of Technology Joint PDF of Paired C, S Samples Relative Probability Lean Aerospace Initiative 180 170 850 160 150 140 130 Schedule (days) Tyson R. Browning — 13 120 550 600 650 700 750 900 800 Cost ($k) PD Browning 101498 ©1998 Massachusetts Institute of Technology Contour Plot View of Joint C, S PDF Lean Aerospace Initiative Budget target 180 Schedule (days) 170 160 150 140 Schedule target 130 120 Tyson R. Browning — 14 550 600 650 700 750 Cost ($k) 800 850 900 PD Browning 101498 ©1998 Massachusetts Institute of Technology One Application: Comparing UCAV Process Configurations Lean Aerospace Initiative .4 .2 .5 .4 .4 .3 .5 .4 .5 .1 .1 .1 .3 .4 .5 .5 .5 .5 .4 .5 Tyson R. Browning — 15 .4 .4 .3 .5 .4 .5 .1 .5 .5 .1 .4 .5 .2 .1 .5 .5 .5 .4 .5 .4 .4 .3 .4 .4 .4 .5 .4 .4 .5 .4 .4 .4 .4 .4 .3 .4 .4 .5 .5 .1 .5 .5 .4 .1 .4 .4 .4 .4 .1 .5 .5 .4 .4 .5 .1 .3 .1 .4 .4 .4 .2 .5 .4 .4 .4 .5 .2 .1 .5 .5 .5 .4 .5 .4 .4 .4 .4 .5 .4 .5 .4 .4 .4 .4 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative Application: Planning a UCAV Process Configuration Linear Scale Cost Risk Plot of RC vs. RS for five process configurations shows trade-off possibilities Tyson R. Browning — 16 PD Browning 101498 ©1998 Massachusetts Institute of Technology Another Application: Choosing Cost and Schedule Targets Lean Aerospace Initiative 1.0 600 0.8 480 0.6 360 0.4 240 0.2 120 0.0 120 0 130 140 150 160 Schedule Target (T Prob. l l l S 170 180 ) Risk Different targets => different levels of risk Choose a target with an acceptable level of risk Other applications: “what if” analysis for adding activities, doing iterations, other process modifications… Tyson R. Browning — 17 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative Managerial Insights l Quantifies relative level of cost and schedule risk in a process configuration (vs. a target) l Compare alternative process configurations and targets for acceptable risk l Visualization of critical path and when activities should wait on each other l Heuristic: minimize long loop iterations in process l Surprising result: prescribe preemptive iteration when appropriate – High set-up time – Low sensitivity to input changes – Significant learning curve effects Tyson R. Browning — 18 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative Relationship Between Activities and Dimensions of Product Performance Activity 1 Cost, Sched. Activity 2 Cost, Sched. Activity 3 Dimension 1 Dimension 2 Cost, Sched. Dimension 3 Activity n Dimension m Cost, Sched. Activities can affect each other Overall Product Performance Product Development Process Determines Performance dimensions can affect each other Premise: By decomposing a process with variance into smaller pieces and understanding the relationships between those pieces, we can better understand the variance of the overall process. Tyson R. Browning — 19 PD Browning 101498 ©1998 Massachusetts Institute of Technology Performance Measure (PM) Tracking Methodology Lean Aerospace Initiative Effective Mission Range PM Forecasting and Tracking 1050 950 Objective 850 Planned 750 650 550 Required Warning Threshold 450 Program Calendar Progress • • • • • Currently used methodology “Events” have effect on PMs Events result from activities Goal of process: to meet requirements and objectives Also: to reduce uncertainty Tyson R. Browning — 20 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative Performance Modeling Methodology l Integrate performance measure (PM) data l Capture how work on activities changes PMs l l l l Separate effects on most likely value (MLV) and uncertainty bounds (decisions vs. tests) Plan project using anticipated effects Manage/Control project using latest evaluations and revised estimates Deterministic and stochastic plans Tyson R. Browning — 21 PD Browning 101498 ©1998 Massachusetts Institute of Technology Types of Activity Effects on PMs Lean Aerospace Initiative Typical Effect on Uncertainty Bounds >< Reign In >+< <><> Uncertain 1 <>+<> <> Increase 2 <+> 3 t1 t2 t1 4 • • t2 t1 <>±<> 5 t2 <±> 6 ? ? • ± Uncertain • ? • • • • ? • • ? • ? ? t1 t2 t1 >–< 7 t2 t1 <>–<> 8 t2 <–> 9 ? • • • • • • ? t1 Tyson R. Browning — 22 • • >±< – Worsen Also, Magnitude Effects Small Medium Large • • ? Typical Effect on Most Likely Value Direction Effects T = set target IE = make initial estimate 1-9 = table: + Improve ? t2 t1 t2 t1 t2 PD Browning 101498 ©1998 Massachusetts Institute of Technology Preliminary UCAV Data: Effects of Activities on Six PMs Lean Aerospace Initiative Magnitude Small Medium Large Direction Performance Measures T = set target Example Set IE = initial estimate 1-9 = table Activities Example Set A4 A511 A512 A531 A521 A522 A5341 A532 A533 A5342 A5343 A5344 A54 A6 Prepare UCAV Preliminary DR&O 1 T T Create UCAV Preliminary Design Configuration 2 5 5 Prepare & Distribute Surfaced Models & Internal Arrangement 3 Drawings Perform Aerodynamics Analyses & Evaluation 4 Create Initial Structural Geometry 5 Prepare Structural Geometry & Notes for FEM 6 Develop Structural Design Conditions 7 Perform Weights & Inertias Analyses & Evaluation 8 7 7 Perform S&C Analyses & Evaluation 9 Develop Balanced Freebody Diagrams & External Applied 10 Loads Establish Internal Load Distributions 11 4 4 Evaluate Structural Strength, Stiffness, & Life 12 4 4 Preliminary Manufacturing Planning & Analyses 13 Prepare UCAV Proposal 14 T T T IE T 5 T 8 T 8 5 8 8 8 T IE, 5 T IE, 5 T 5 8 7 7 T Mapping the typical, direct effects (direction and magnitude) of activities on PMs Tyson R. Browning — 23 PD Browning 101498 ©1998 Massachusetts Institute of Technology Stochastic Model: Sample Result Lean Aerospace Initiative 1000 MLV E[V] 800 600 400 200 0 0 20 40 60 80 100 120 140 160 Project Time (t, days) l l l l Simulated PM tracking chart for UCAV payload Changes in MLV and uncertainty caused by activities in process Goals of process are improved MLV and reduced uncertainty Current limitation: no feedback from current performance level Tyson R. Browning — 24 PD Browning 101498 ©1998 Massachusetts Institute of Technology Preliminary, Example Result (Single Run) Lean Aerospace Initiative (Revise) Configuration Good Struct. Str. & Life Good Internal Loads RP Good Struct. Str. & Life Structural Geometry Affects Reliability 160 140 Fourth Weights Analysis 120 100 80 60 40 20 0 Second Weights Analysis Mfg. Planning • Compare processes • Add activities • Repeat activities • Repeat processes Project Time (t, days) Prepare UCAV Preliminary DR&O Create UCAV Preliminary Design Configuration Prepare & Distribute Surfaced Models & Internal Arrangement Drawings Perform Aerodynamics Analyses & Evaluation Create Initial Structural Geometry Prepare Structural Geometry & Notes for FEM Develop Structural Design Conditions Perform Weights & Inertias Analyses & Evaluation Perform S&C Analyses & Evaluation Develop Balanced Freebody Diagrams & External Applied Loads Establish Internal Load Distributions Evaluate Structural Strength, Stiffness, & Life Preliminary Manufacturing Planning & Analyses Prepare UCAV Proposal Tyson R. Browning — 25 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative l Options and Applications Options: – Repeat (iterate) all or portion of process – Add new activities to affect PMs l Applications: – Exploring CS&P implications of options – Exploring PM contributions to RP – Exploring activity contributions to RP – Ensuring existence of risk reducing (value adding) chain of activities – Comparing processes l Result: – Maintained balance and congruence of product and process to understood customer desires Tyson R. Browning — 26 PD Browning 101498 ©1998 Massachusetts Institute of Technology Lean Aerospace Initiative l l l l l Managerial Insights Process cost and schedule affect product performance Compare processes to see which configurations maximize overall value to customer (best balance of C,S, and technical P) Explore effect of management decisions on RP Check integrity of value stream (i.e., all necessary, value-adding activities) And more… Tyson R. Browning — 27 PD Browning 101498 ©1998 Massachusetts Institute of Technology