Value-driven design

Abstract: The inability to approach systematically the high level of ambiguity present in the early design phases of space systems causes long, highly iterative, and costly design cycles. A process is introduced and described to capture decision maker preferences and use them to generate and evaluate a multitude of space system designs, while providing a common metric that can be easily communicated throughout the design enterprise. Communication channeled through formal utility interviews and analysis enables engineers to better understand the key drivers for the system and allows for a more thorough exploration of the design tradespace. Multi-attribute tradespace exploration with concurrent design, a process incorporating decision theory into model- and simulation-based design, has been applied to several space system projects at the Massachusetts Institute of Technology. Preliminary results indicate that this process can improve the quality of communication to resolve more quickly project ambiguity and to enable the engineer to discover better value designs for multiple stakeholders. The process is also integrated into a concurrent design environment to facilitate the transfer of knowledge of important drivers into higher fidelity design phases. Formal utility theory provides a mechanism to bridge the language barrier between experts of different backgrounds and differing needs, for example, scientists, engineers, managers, etc. Multi-attribute tradespace exploration with concurrent design couples decision makers more closely to the design and, most important, maintains their presence between formal reviews.

Abstract: Engineering design needs to employ the necessary multidisciplinary skills and tools to generate the alternative designs and identify the optimal one. To perform the most successful design optimization, the needs and priorities of all stakeholders involved with the engineering design should be considered. Towards this goal, a standardized process is presented, implementing the Value Driven Design (VDD) philosophy and allowing to perform a multi-objective and multi-stakeholder engineering design optimization. The needs of all major stakeholders involved with the designed system are captured with Game Theory, forming a hybrid cooperative/non-cooperative, non-zero sum, complete information game. In this hybrid game, the classical cooperative games are combined with the non-cooperative game of Game Theory to identify the optimal engineering design that has the property of being the Nash bargaining solution (NBS) simultaneously, capturing both the conflict and the synergy between the stakeholders. Thus, engineering design is converted to a formalized decision-making process, taking into account all the objectives of the stakeholders involved with the designed system.

Abstract: : One of the most ambitious efforts in value-centric design of a military aerospace system undertaken to date has been the parallel development by four performer teams, headlined by major space industry primes, of design tools for fractionated space architectures under DARPA's System F6 program. The goal of the System F6 program is to replace traditional, highly-integrated, monolithic satellites with wirelessly-networked clusters of heterogeneous modules incorporating the various payload and infrastructure functions. Such fractionated architectures can deliver a comparable or greater mission capability than monolithic satellites, but with significantly enhanced flexibility and robustness. In order to design an optimal fractionated architecture, the potential cost penalties due to the overhead of such a design must be balanced against the value enhancement due to improved flexibility and robustness. The first, preliminary design phase of the System F6 program, simultaneously awarded to four competing industry teams led by Boeing, Lockheed Martin, Northrop Grumman, and Orbital Sciences, commenced in February 2008 and included a significant effort for the development, validation, and demonstration of a Value-Centric Design methodology and associated tool suite that can support the design of optimized fractionated satellite systems based on a net lifecycle value metric and a probabilistic distribution thereof. This phase concluded in February 2009 and the Value-Centric Design methodology development to date is documented in a series of papers by the industry performer teams. This paper, from the System F6 Program Office, summarizes the overarching objectives of the Value-Centric Design effort, details and rationalizes the requirements for the methodology, discusses the relationship between Value-Centric Design and the traditional industry-standard systems engineering process, and fills any gaps in the performers' own presentations of their efforts, tools, and results.

Abstract: Spacecraft * and launch systems are examples of complex products which require a careful balance between competing concerns, such as performance, weight, and reliability, to serve their mission. Complexity requires design by large engineering organizations, so this balance must be achieved across many teams of people working on various components. This paper uses optimization theory to derive a method for distributed optimal design. Each component design team is provided with a separate optimization problem such that, as each team finds the best design solution to their problem, the teams together design the best system. To date, distributed optimal design has been difficult because complex system design spaces have extremely high dimensions over which design objectives are poorly correlated. Instead, the paper proposes that design objectives be expressed as functions in property spaces, which have few dimensions and are much smoother than design spaces. Property spaces are generated from design spaces by traditional engineering analysis processes. Economic analysis of all parties to the spacecraft launch and operation is used to construct a top-level value function on the system property space. This function is linearly decomposed into value functions for component property spaces. This provides the needed objective functions for distributed optimal design.