LAV25 Logistics Optimization using Machine Learning

Risk-based time-to-event (TTE) modeling across the LAV fleet will be used to improve strategic mission-critical MAGTF scenario planning. This project seeks to model, quantify and forecast LAV operational availability as a function of time. Consolidated key intelligence data is used to estimate the probability of an event occurring in the future (given all past maintenance related information) and predict the associated time-to-event as a function of maintenance status (e.g. deadlined, operational-degraded, etc.), cost and failure mode (parts). By integrating service request records and supply activities at scale and leveraging complex computational principles, Tagup has developed methods of extracting and structuring large volumes of existing data while creating methods to predict asset availability outcomes more reliably. This modeling/technical approach leverages a wealth of maintenance and supply data in the existing Global Combat Support System (GCSS-MC) database [rated Controlled Unclassified Information (CUI)] therefore mitigating installation of expensive sensor equipment. During Phase I, Tagup developed and evaluated ML based techniques that could be used to forecast LAV availability as a function of historical use, health and reliability (all captured in GCSS-MC). Tagup’s technical approach successfully demonstrated a means to improve Marine Air-Ground Task Force’s (MAGTF) scenario planning capability (e.g. deployment to a remote desert environment), by: Providing a means to increase operational availability of the LAV through a reduction in short parts, Logistics Response Time (LRT) and overall Customer Wait Time (CWT) Providing a method to increase confidence in asset readiness and inventory posture Through a series of three (3) in-person meetings held during Phase I with key USMC stakeholders (LOGCOM and SYSCOM), Tagup confirmed the primary research objectives, analytic needs, product capabilities and proposed methods to advance the research developed during Phase I for productization, implementation and validation in Phase II. Phase II will operationalize and validate these analytic methods as defined in the Technical Objectives. In order to validate the technical approach and related benefits, a continuation to Phase II would allow Tagup to continue working with MARCORSYSCOM and MARCORLOGCOM to: Build and deploy two (2) new simulation/forecasting tools on streaming GCSS-MC data (as confirmed by the technical approach demonstrated in Phase I) OPerational Readiness & future Availability (OPRA) Demand-Driven LOgistics Simulation Tool (LOST) Validate and operationalize the analytics through block building, deployment planning and maintenance and supply optimization Certify Tagup’s software meets DOD cybersecurity requirements Savings realized due to increased asset availability will be quantified in detail. These savings will support commercialization plans across additional TAMCNs and users.