Accession Number : ADA545026


Title :   Efficient Effects-Based Military Planning Final Report


Descriptive Note : Final rept. 1 Aug 2006-31 Jul 2010


Corporate Author : RENSSELAER POLYTECHNIC INST TROY NY DEPT OF ELECTRICAL COMPUTER AND SYSTEMS ENGINEERING


Personal Author(s) : Ji, Qiang


Full Text : https://apps.dtic.mil/dtic/tr/fulltext/u2/a545026.pdf


Report Date : 13 Nov 2010


Pagination or Media Count : 179


Abstract : This research focused on developing a mathematical framework and the associated methods for effects-based military plan modeling and evaluation. Specifically, we first introduced a unified probabilistic framework based on the Dynamic Influence Diagram to systematically represent the causal relationships between actions and their effects, their interactions, their uncertainties, and their dynamics. We then developed advanced machine learning methods to automatically construct such a model for a given campaign and to learn the model parameters from both the training data and the available qualitative domain knowledge. Given the framework, we developed efficient inference methods to quickly evaluate each plan to identify the optimal plan. This research also includes the development of a user-friendly prototype software to evaluate the performance of the proposed methods for military plan modeling and evaluation. The software can produce a model for a military campaign, automatically learn the model structure and parameters, evaluate various possible plans, and identify the one that best meets the campaign objective.


Descriptors :   *MILITARY PLANNING , *SOFTWARE TOOLS , REASONING , LEARNING MACHINES , USER FRIENDLY , CYBERNETICS , PARAMETERS , DECISION MAKING , ARTIFICIAL INTELLIGENCE


Subject Categories : Information Science
      Cybernetics
      Military Operations, Strategy and Tactics


Distribution Statement : APPROVED FOR PUBLIC RELEASE