System and method for modeling and analyzing complex scenarios

• US 8,412,708 B2
• Filed: 09/17/2009
• Issued: 04/02/2013
• Est. Priority Date: 09/17/2009
• Status: Active Grant

• Current Assignee: TRIAD National Security, LLC
• Original Assignee: Los Alamos National Security LLC (Government of the United States of America)
• Inventors: Shevitz, Daniel Wolf
• Primary Examiner(s): Breene, John E
• Assistant Examiner(s): MYINT, DENNIS Y
• Application Number:
  US12/561,729
• Publication Number:
  US 20110066621A1
• Time in Patent Office: 1,293 Days
• Field of Search: 707/736, 707/797, 707/803, 707/804
• US Class Current: 707/736
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation

Abstract

An embodiment of the present invention includes a method for analyzing and solving possibility tree. A possibility tree having a plurality of programmable nodes is constructed and solved with a solver module executed by a processor element. The solver module executes the programming of said nodes, and tracks the state of at least a variable through a branch. When a variable of said branch is out of tolerance with a parameter, the solver disables remaining nodes of the branch and marks the branch as an invalid solution. The valid solutions are then aggregated and displayed as valid tree solutions.

Claims

1. A method of possibility tree analysis, the method comprising:

constructing a possibility tree having a plurality of nodes, wherein at least one said node being programmable, and populating said at least a programmable node with at least a variable;

programming at least a said programmable node with instructions for utilizing said at least a variable during the execution of said possibility tree;

solving said possibility tree, where said solving is performed by a solver module executed by a processor element, and where said solving includes;

tracking a path state dictionary of variables specific to a path of said possibility tree along said path, where a said programmable node creates and adds a variable to said path state dictionary, and modifies a state of an existing said variable in said path state dictionary,

comparing at least a said variable to a user defined condition as instructed within the programming of a said programmable node,

disabling remaining nodes of said path during the execution of said path and marking said path as an invalid solution when a said variable of said path meets a user defined condition,

repeating said tracking, comparing and disabling steps for a plurality of paths, and aggregating and displaying completed paths as valid tree solutions,

wherein the construction of the possibility tree includes creating and placing nodes within the possibility tree, said creating and placing including the steps of;

creating a gate information module containing instructions for executing a node, said instructions including programming;

creating at least two context information modules, where each context information module contains information defining the location within the possibility tree of a respective node;

associating the at least two context information modules with the gate information module, thereby creating at least two replicant nodes.

2. The method of claim 1 wherein said populating of said at least a programmable node includes programming said programmable node with database lookups, accessing a database according to said database lookups, and importing at least a said variable according to said database lookups for use in the possibility tree.

3. The method of claim 1 where, upon reaching the conclusion of a path, the possibility tree is rolled back to its state as of its last OR gate node with an unused option, allowing the execution of said possibility tree to continue with the selection of the unused option.