Using Big Data Analytics to Create a Predictive Model for Joint Strike Fighter
The amount of information needed to acquire knowledge on today‚??s acquisition systems is growing exponentially due to more complex, higher resolution, software-intensive acquisition systems that need to operate in System-of-Systems (SoS), Family-of-Systems (FoS), Joint, and Coalition environments. Unfortunately, the tools and methods necessary to rapidly collect, aggregate, and analyze this information have not evolved as a whole in conjunction with this increased system complexity and, therefore, has made analysis and evaluation increasingly deficient and ineffective. The Test Resource Management Center‚??s (TRMC‚??s) vision is to build a DoD test and evaluation (T&E) knowledge management (KM) and analysis capability that leverages commercial big data analysis and cloud computing technologies to improve evaluation quality and reduce decision-making time. An evaluation revolution, starting with the Joint Strike Fighter (JSF) program, is underway to ensure the T&E community can support the demands of next-generation weapon systems. The true product of T&E is knowledge ascertained through the collection of information about a system or item under test. However, the T&E community‚??s ability to provide this knowledge is hampered by more complex systems, more complex environments, and the need to be more agile in support of strategic initiatives, such as agile acquisition and the 3rd Offset Strategy. This increased complexity and need for speed cause delayed analysis and problems that go undetected during T&E. The primary reason for these shortfalls is antiquated tools and processes that make data hard to locate, aggregate, and convert into knowledge. In short, DoD has not evolved its evaluation infrastructure as its weapon systems have evolved. Conversely, commercial entities, such as medical observation and diagnosis, electric power distribution, retail, and industrial manufacturing, have embraced agility in their methodologies while modernizing analytics capabilities to keep up with the massive influx of data. Raw physical sensors could provide data, higher-quality image or video cameras, radio frequency identification (RFID) devices, faster data collectors, more detailed point-of-sale information or digitized records, and ultimately is providing more data to analysts in size and complexity than ever before. As more data has become available, an interrelated phenomenon is the desire of analysts to ask more detailed questions about their consumers and their business infrastructure. To drive the process of implementing big data analytics, businesses have begun establishing analytics centers which either take pre-defined business cases and apply methods to address them or implement existing knowledge within the data architecture to create a higher level of awareness to business groups or the company at-large. To meet these demands, data storage and computation architectures have become more sophisticated, dozens of technologies were developed for large-scale processing (such as Apache Hadoop or GreenPlum), and streaming architectures which allow data to be processed and actioned on in real-time as it is collected have become commonplace. The net result of these commercial best practices is a solid foundation for the DoD to transform how it uses data to achieve faster, better, and smarter decisions throughout the acquisition lifecycle.
Big Data, Data Analytics, Knowledge Management, Data Management, Virtualization, Cloud Computing, Predictive Maintainance, Department of Defense, Test and Evaluation