Combining Heuristic Search, Visualization and Data Mining for Exploration of System Design Spaces

摘要

Finding a near-optimal solution to a complex design problem is challenging. Onrnthe one hand the problem space is too large and convoluted for human comprehension, while onrnthe other hand it is infeasible to elicit the entirety of design knowledge required for fullyrnautomatic problem solving.rnWe report on a practical application of information technology techniques to aid systemrnengineers effectively explore large design spaces. We make use of heuristic search, visualizationrnand data mining, the combination of which we have implemented within a risk management toolrnin use at JPL and NASA.rnThis approach is demonstrated on the planning for development of an advanced technologyrnfor spacecraft applications. In this context of risk-informed design, numerous risk abatementrnoptions give rise to a huge space of potential design solutions. We show how our approachrnenhances the system engineers’ ability to explore this design space.