SBIR-STTR Award

In a pilot study effort in 2013, the Secretary of Defense for Manufacturing and Industrial Base Policy (ODASD (MIBP)) developed a methodology that goes beyond legacy reactive and program-centric frameworks for assessing industrial base risk. This proposed effort leverages the pilot studys work with the fragility and criticality (FaC) assessment to develop a predictive and proactive tool to assist in the analysis and categorization of industrial base risks. The risk mitigation tool will use advanced data warehouses to ingest identified sources of relevant data and will employ unsupervised deep learning algorithms to identify fragility and criticality patterns in the dataset. Other untapped sources of data will be automatically gathered by the tool to assess industrial base risk. This innovative way of predicting supply risks will be extremely efficient and accurate and will require less direct human interaction than the traditional FaC process. UTC/UDRI team includes a commercialization partner, capable of providing FASI-G (Fleet Automotive Support Initiative - Global) data and transitioning this technology to several targeted Department of Defense (DoD) platforms. Our technical team includes members with relevant experiences in advanced machine learning algorithms and data analytics. Approved for Public Release | 17-MDA-9219 (31 May 17)

Presently the missile defense systems are using a reactive and program-centric framework for assessing industrial base risk. The Phase I effort focused on developing a wide and deep network algorithm using Industrial Product-Support Vendor (IPV) Gen II program for predicting the probability of failure. The proposed Phase II effort leverages the Phase I work and will focus on three main goals:(1) Inclusion of Bayesian networks to the existing wide and deep network algorithm to extract risk correlations;(2) Implementing a secure cloud based module for ingesting new data sources - weather, financial and event data;(3) Implementing an application that integrates the algorithmic and data sourcing components in an adaptive framework that can predict the likelihood of supply chain failure and identify the key risk factors that create those failures. Phase II will result in a software program that will adaptively gather relevant industrial supply data and provide an almost real time prediction of industrial risks. UTC/UDRI team includes a commercialization partner, Lockheed-Martin (LM), capable of providing data and transitioning this technology to several targeted DoD platforms. Our team includes members with relevant experience in advanced machine learning algorithms, data analytics, and supply chain management.Approved for Public Release | 18-MDA-9522 (23 Feb 18)