Genetically Trained Fuzzy Cognitive Maps for Effects Based Operations

Effects Based Operations (EBO) is an umbrella term for operational environments in which an action or series of actions cascade through chain-effects to produce notable changes in some key end-chain effects. Important examples are wartime theaters, city planning, politics, and advertising. The key component of EBO analysis is determining the best set of actions for maximization of a desired effect or minimization of an undesired effect. EBOs are analyzed using knowledge graphs. Fuzzy Cognitive Maps (FCM) are one method of analytically representing and solving knowledge graphs. Integrating a fuzzy inference system on each node of a knowledge graph allows for the traditional relational weight matrix to be replaced with fuzzy rule sets allowing for more uncertainty in relational effect weights. This reflects real world uncertainty in wartime, political, or other planning environments. Using a genetic algorithm also decreases the vast search space of possible knowledge graph solutions. This allows for a robust analysis of an FCM-EBO system with a high level of explainability retained in the system.