Decision Support

Intelligent decision support systems help people assess situations, analyze information, and make predictions or decisions in areas such as equipment diagnosis, planning, and design. These systems apply the knowledge of experts, encoded as heuristic rules, computational models, or experiences.

Success Stories

 

  • Measure and Visualize Source Code Quality
    CBR Insight (CBRI) provides an objective and understandable measure of software quality that can help guide decisions during software acquisition, development, and sustainment. CBR Insight 1) examines source code to calculate metrics that are highly related to software reliability, maintainability, and preventable technical debt, 2) develops realistic target ranges for these metrics based on successful ‘peer’ projects, 3) generates an aggregated score by comparing metric calculations to the target ranges, and 4) presents an accessible dashboard overview of results to decision makers. Datasheet
  • Team Environment Analysis and Modeling
    TEAM enables organizations to understand, analyze, and track inter-related political, military, economic, social, and technological factors to help them determine how best to influence their environment and achieve their objectives.
  • Improving Surface Threat Detection and Identification
    Stottler Henke developed a software system, the Intelligent Surface Threat Identification System (ISTIS), based on Artificial Intelligence (AI) techniques that improves the surface threat ID process, quality, and efficiency in the Littoral Combat Ship and its Surface Mission Module.
  • A Beacon for Military Plan Execution and Outcomes
    Stottler Henke is developing BEACON, a system that provides an ontology of concepts suitable for representing key aspects of military plans and the narrative structure of historical cases that shed light on plan execution and outcomes. BEACON contains a corpus of cases, constructed using semi-automated methods, intended to mimic the ideal results of advanced information extraction technology. BEACON contains algorithms for indexing those cases for rapid retrieval, matching in more detail to determine case relevance, and presenting retrieved cases, including their relevance to plan-related queries. The user interface leverages earlier Stottler Henke work on the TEAM campaign design and planning wiki, intended as a surrogate for capabilities of an advanced collaborative design and planning support tool.
  • Improved SSN Scheduling
    We have developed a scheduling algorithm that takes as input the space catalog and the associated covariance matrices and produces a globally optimized schedule for each sensor site as to what objects to observe and when. This algorithm is able to schedule more observations with the same sensor resources and have those observations be more complementary, in terms of the precision with which each orbit metric is known, to produce a satellite observation schedule that, when executed, minimizes the covariances across the entire space object catalog.

  • Applying Deep Learning to Improve Maritime Situational Awareness
    Stottler Henke developed an Extensible Platform for Automated Tactical Sensor Screening (ExPATSS) for the Navy to automatically detect and classify ships by simultaneously processing several video streams and applying deep learning to perform ship detection and classification.

  • Predicting Retail Sales
    Stottler Henke developed a case-based reasoning (CBR) system that predicts the daily sales volume for each store in a store chain of 800 service centers to support staffing decisions that ensure adequate yet efficient staffing levels.