Brown, Rockwell, University of Washington TIE - Roadmap

Here's what we currently intend regarding the Brown, Rockwell, and University of Washington (BRW) TIE.

Anticipated Results

First, the proposed extension will provide operators with the ability to construct more efficient and robust plans based on expected rather than worst-case outcomes. Second, the extension will provide an opportunity to add value by automating aspects of model formulation and source generation. Operators can use these added-value capabilities as an advanced form of evaluation, adopting suggestions only if they are confident in the program.

Agenda

  1. Obtain the CTEMS linear program.
  2. Obtain and install ACPT from ISX.
  3. Learn about ACPT. In particular, find out how to extract the necessary information for quantifying a probabilistic model from the preliminary assessment phase in ACP.
  4. Add a window (tool) that takes CTEMS output (a sourcing) and evaluates it as a plan (degenerate policy) using the probabilisitc model mentioned in (3).
  5. Add value to the tool by providing simple sensitivity analysis (sliders to play with attrition rates and damage assessments).
  6. Add even more value, ultimately replacing CTEMS, by generating real policies with conditional sourcing.
  7. Iterate on the design of 4-6 with feedback from CHECKMATE.
(5-7) currently are not part of the TIE, i.e., we're not willing to commit to these items at this stage. We are interested in these additions but there is too much uncertainty in the levels of cooperation from ACP to make any reasonable predictions at this time.

Evaluation

We intend to enter negotiations with the Air Force Checkmate Division regarding evaluation methods as soon as the relationship with Rome Labs and the Planning Initiative is formalized.