The Tradespace Exploration Paradigm Adam Ross and Daniel Hastings MIT INCOSE International Symposium July 14, 2005

Transcription

1 The Tradespace Exploration Paradigm Adam Ross and Daniel Hastings MIT INCOSE International Symposium July 14, 2005

2 2of 17 Motivation Conceptual Design is a high leverage phase in system development Need Captured 100% 80% ~66% 66% Lifecycle LCC Cost committed Cost Incurred Design Resources Scoped Ease of Change N EE D Conceptual/ preliminary Design Detail design/ development Production and/or construction Product use/ support/ phaseout/disposal from Fabrycky, 1991 In Situ vs. Top-side sounder Concept Selected

3 3of 17 Glossary Attribute. A decision-maker perceived metric that measures how well a decision maker-defined objective is met Decision Maker. A type of stakeholder that has significant influence over defining system objectives or allocation of resources (DM) Design Variables. Designer controlled knobs that represent aspects of the system concept -CON. Multi-Attribute Tradespace Exploration with Concurrent Design couples broad tradespace exploration with explicit decision maker value functions Tradespace. The space spanned by the completely enumerated design variable set, often represented by (cost, utility) per DM Utility. Dimensionless parameter that reflects the perceived value under uncertainty of an attribute; a rational DM seeks to maximize utility Value. Like Beauty, perception of goodness (value) is often subjective

4 4of 17 Avoiding Point Designs Differing types of trades Utility 1. Local point solution trades 2. Frontier subset solutions 3. Frontier solution set 4. Full tradespace exploration Design i = {X 1, X 2, X 3,,X j } Cost Tradespace exploration enables big picture understanding

5 5of 17 Multi-Attribute Tradespace Exploration 1 Context: Engineering Systems Thinking 2 Inputs 3 Model/ Simulation Outputs 4 Focus of talk Value-centric Design Model/Sim Tradespace Analysis Techniques Decision Theory Design Theory Lean Research Parametric Models Dynamic Models Integrated Concurrent Design Similarity analysis Sensitivity analyses Portfolio Theory Real options

6 6of 17 What is an Architecture Tradespace? Stakeholders Value Attributes Utility Analysis Concept Design Cost Tradespace: {Design,Attributes} {Cost,Utility} Each point is a specific architecture X-TOS Small low-altitude science mission km Km DESIGN VARIABLES: Architectural trade parameters Orbital Parameters Apogee Altitude (km) Perigee Altitude (km) Orbit Inclination (deg) Spacecraft Parameters Antenna Gain Communication Architecture Propulsion Type Power Type Total Delta V Cost, Utility Total Lifecycle Cost ($M2002) ATTRIBUTES: Architectural decision metrics Data Lifespan (yrs) Equatorial Time (hrs/day) Latency (hrs) Latitude Diversity (deg) Sample Altitude (km) Assessment of cost and utility of large space of possible system architectures

7 7of 17 : Rqmts Easily Assessed Weight Factors of each Attribute (k values) Architecture tradespace reevaluated in less than one hour 0 Original Latency Latitude Equator Revised Time Lifespan Altitude Original Revised User changed preference weighting for lifespan X-TOS Industry, Government From Ross, and 2003 Academia

8 8of 17 : Understanding Limiting Physical or Mission Constraints Cost ($M) Spacetug Tradespace Low Biprop Prop Type Medium SPACETUG Biprop Biprop High Biprop Cryo General purpose orbit Extreme Biprop Electric transfer Low Cryo vehicles Nuclear Medium Cryo Different propulsion High Cryo systems Extreme Cryoand grappling/ Low Electric Medium Electric observation High Electric capabilities Extreme Electric Low Nuclear Lines show increasing Medium Nuclear fuel High mass Nuclear fraction Extreme Nuclear See McManus and Schuman, 2003 Utility (dimensionless) Hits wall of either physics (can t change!) or utility (can)

9 9of 17 Tradespace Exploration w/ Uncertainty 500 Cost B Architectures: Changes (in anything) may cause large added cost A Architectures: Changes (in anything) have less drastic affect; more value may be available for modest added cost Utility (dimensionless) Often learn a lot by simple examination Better: Explicitly look at model sensitivities to uncertainties Uncertainties can be market (shown), policy, or technical Mitigate with portfolio, real options methods From Walton, 2002

10 10 of 17 Portfolio Analysis: 3 DM s Percentage of Portfolio 52% 48% 100% low high Uncertainty in the High Risk Aversion Portfolio is less than each of its assets Risk covered by investment in small system that only does one mission A portfolio is investment in multiple designs Low Risk Aversion Portfolio contains Best Value Design (I.e. Highest TU/$) Moderate Risk Aversion High Risk Aversion Low Risk Aversion Percentage Architecture Design Vector Total Uncertainty If designs are of Portfolio anticorrelated {sats/swarm,suborbs,size,yaw,subplaces,alt} Utility/$ with respect Architecture Design Vector Total Uncertainty Percentage Architecture Design Vector {sats/swarm,suborbs,size,yaw,subplaces,alt} Utility/$ 57% {26,4,14.1,60,2,700} of Portfolio {sats/swarm,suborbs,size,yaw,subplaces,alt} {26,4,14.1,60,2,700} % {4,2,3.8,30,1,500} % {8,4,14.1,30,1,700} to uncertainties, portfolios can have lower {2,1,3.8,30,1,300} % {4,1,14.1,0,1,700} % {4,2,3.8,30,1,500} Portfolio Value and Uncertainty % Portfolio Value and Uncertainty % Portfolio Value and Uncertainty uncertainty than individual designs Total Utility/$ Uncertainty Optimal Strategy Portfolio Optimal Strategy Portfolio Optimal Strategy Portfolio From Walton, 2002

11 11 of 17 Flexibility Over Time: SBR Technology and Programmatic Flexibility Over Time for Space-Based Radar (SBR) Utility Number of Transition Possibilities: 2002 to 2005 (decreasing altitude) Cost of flexibility? Life cycle Cost ($B) Additional tech transitions possible based on investment From Shah, 2004 Transitionability relates to distance between architectures Transition costs vary widely based on transition path Optimality not defined when fitness function changes over time Pareto Front may not provide best answers

12 12 of 17 Application: Spiral Development Evolution of Capability in Second Spiral for Small Diameter Bomb Exp MAU Tradespace ParetoFront First Spiral Pareto Set Exp Cost ($) 9 10 x 10 8 From Derleth, 2003 Pareto set differentiates over time Marginal cost versus rank statistics reveal best baselines System concept is a small, airplane-carried bomb Spiral two adds new attributes to utility function Tradespace gives insights in planning for spiral development

13 13 of 17 Application: Comparing Point Designs Designs from traditional process TPF Terrestrial Planet Finder - a large astronomy system Design space: Apertures separated or connected, 2- D/3-D, sizes, orbits Images vs. cost From Jilla, 2002 Existence of dominated point designs focus discussion Industry, Government [Beichman and et Academia al, 1999]

14 14 of 17 Application: Budget Cap Policy Utility Policy Intervention: $35M Annual Program Budget Cap Imposed by Congress Policy results in Key: Nominal architecture differential cost Pareto front, nominal architectures increases Cost-capped budget architecture Pareto front, costcapped architectures Policy-robust points remain in Pareto set Lifecycle Cost ($M) From Weigel, 2002 System concept is a satellite swarm that samples Earth s ionosphere Cost-capping policy pushes Pareto Front to the right

15 15 of 17 Application: Multi-DM Negotiation DM1 Multiple Decision Maker Tradespace 1 DM2 Pareto surface 2 In general, no optimum solution to problem with multiple stakeholders having conflicting needs (See Arrow s Impossibility Theorem in Hazelrigg, 1996) Tradespace aids negotiation 1. Finding the win-win changes (moving toward Pareto front) 2. Finding the real trades (along the Pareto front) Negotiation advantage for stakeholders who understand tradespace Explicit value conflict and congruity become apparent in tradespace

16 16 of 17 Tradespace exploration is a unified framework that enables Consideration of diverse and dynamic value functions Comparison of diverse and dynamic concepts Characterization and mitigation of various uncertainties Quantification of system properties (e.g., flexibility, robustness) have included: policy sensitivity analysis, spiral development, cross-proposal evaluation On-going research seeks to standardize TSE, including theory, method and applications

17 17 of 17 Citations Beichman, C.A., Woolf, N.J., and Lindensmith, C.A., "The Terrestrial Planet Finder (TPF): A NASA Origins Program to Search for Habitable Planets," JPL Publication 99-3, May 1999, pp. 1-11, 49-55, Derleth, Jason E. "Multi-Attribute Tradespace Exploration and Its Application to Evolutionary Acquisition." SM, Massachusetts Institute of Technology, Fabrycky, W.J. Life Cycle Cost and Economic Analysis. Prentice-Hall, NJ Hazelrigg, George A. Systems Engineering: An Approach to Information-based Design. Upper Saddle River, NJ: Prentice Hall, Jilla, Cyrus D. "A Multiobjective, Multidisciplinary Design Optimization Methodology for the Conceptual Design of Distributed Satellite Systems." Ph.D., Massachusetts Institute of Technology, McManus, H. and T. E. Schuman. Understanding the Orbital Transfer Vehicle Trade Space. AIAA Space 2003 Conference and Exhibition, Long Beach, CA, Ross, Adam M. "Multi-Attribute Tradespace Exploration with Concurrent Design as a Value-Centric Framework for Space System Architecture and Design." Dual-SM, Massachusetts Institute of Technology, Shah, Nirav B. "Modularity as an Enabler for Evolutionary Acquisition." SM, Massachusetts Institute of Technology, Walton, Myles. "Managing Uncertainty in Space Systems Conceptual Design Using Portfolio Theory." PhD, Massachusetts Institute of Technology, Weigel, Annalisa L. "Bringing Policy into Space Systems Conceptual Design: Quantitative and Qualitative Methods." PhD, Massachusetts Institute of Technology, 2002.

18 More Sources

19 19 of 17 References (1) (8 MS theses) Derleth, Jason E. "Multi-Attribute Tradespace Exploration and Its Application to Evolutionary Acquisition." SM, Massachusetts Institute of Technology, Diller, Nathan P. "Utilizing Multiple Attribute Tradespace Exploration with Concurrent Design for Creating Aerospace Systems Requirements." SM, Massachusetts Institute of Technology, Roberts, Christoper J. "Architecting Strategies Using Spiral Development for Space Based Radar." SM, Massachusetts Institute of Technology, Ross, Adam M. "Multi-Attribute Tradespace Exploration with Concurrent Design as a Value-Centric Framework for Space System Architecture and Design." Dual-SM, Massachusetts Institute of Technology, Seshasai, Satwiksai. "A Knowledge Based Approach to Facilitate Engineering Design." M.Eng., Massachusetts Institute of Technology, Shah, Nirav B. "Modularity as an Enabler for Evolutionary Acquisition." SM, Massachusetts Institute of Technology, Spaulding, Timothy J. "Tools for Evolutionary Acquisition: A Study of Multi-Attribute Tradespace Exploration () Applied to the Space Based Radar (SBR)." SM, Massachusetts Institute of Technology, Stagney, David B. "The Integrated Concurrent Enterprise." SM, Massachusetts Institute of Technology, 2003.

20 20 of 17 References (2) -related (3 PhD dissertations) Jilla, Cyrus D. "A Multiobjective, Multidisciplinary Design Optimization Methodology for the Conceptual Design of Distributed Satellite Systems." Ph.D., Massachusetts Institute of Technology, Walton, Myles. "Managing Uncertainty in Space Systems Conceptual Design Using Portfolio Theory." PhD, Massachusetts Institute of Technology, Weigel, Annalisa L. "Bringing Policy into Space Systems Conceptual Design: Quantitative and Qualitative Methods." PhD, Massachusetts Institute of Technology, Precursor theses Browning, Tyson R. "Modeling and Analyzing Cost, Schedule, and Performance in Complex System Produce Development." PhD, Massachusetts Institute of Technology, Delquie, Philippe. "Contingent Weighting of the Response Dimension in Preference Matching." Ph.D., Massachusetts Institute of Technology, Nolet, Simon. "Development of a Design Environment for Integrated Concurrent Engineering in Academia." M. Eng., Massachusetts Institute of Technology, Shaw, Graeme B. "The Generalized Information Network Analysis Methodology for Distributed Satellite Systems." Sc.D., Massachusetts Institute of Technology, 1999.