اکتشاف بسیار کارآمد از فضاهای طراحی بزرگ: ماهواره های مختلط به عنوان مثال از سیستم های قابل تنظیم

The design of systems with complex interdependencies requires evaluation of a vast number of potential designs against performance and cost based metrics. For example, the Defense Advanced Research Projects Agency's (DARPA's) System F6 (Future Fast, Flexible, Fractionated, Free-Flying Spacecraft united by Information eXchange) program aims to demonstrate the feasibility of satellite fractionation. In a major departure from traditional “monolithic” satellites, an F6 system is a cluster of heterogeneous, free-flying, wirelessly interconnected modules that evolves over time to provide better robustness to failures and flexibility to changing needs and improving technologies. The cost and value of any individual spacecraft in an F6 system can only be estimated in the context of the whole constellation and its evolution. This creates new challenges in spacecraft design, but also new opportunities. We present the use of ISI's SPIDR (System Platform for Integrated Design in Real-time), an artificial intelligence-based search and optimization engine, in tackling fractionation's vast trade space. Originally developed for design of monolithic spacecraft, here SPIDR is used to design an evolving F6 constellation. We discuss our approach to modeling sharing and evolution in an adaptable space system and present examples of variations in optimal designs as system constraints and optimization metrics vary.