Intelligent Hierarchical Decision Architecture for Operational Test andEvaluation

摘要

Operational Test and Evaluation (OTE) is testing conducted on a full-up system,to determine for a decision-maker, if it will be able to perform the tasks for which it was designed. Current methods used for the analysis of OTE data include standard statistical methods that are adequate for summarizing the information gathered during a testing phase (information at a functional-performance level). However, these methods have proved inadequate in providing the type of information needed by the decision-maker in the OTE context: information on the system's ability to accomplish the tasks for which it was designed. The Intelligent Hierarchical Decision Architecture provides a neuro-fuzzy analysis methodology meant to take test data at the functional-performance level gathered in a laboratory, through modeling and simulation, or through field testing and aggregates/synthesizes it to provide information at the task-accomplishment level, where it is meaningful to the decision-maker. The analysis methodology is composed of four separate stages. Raw test data enters the Intelligent Hierarchical Decision Architecture as individual observations of the system's functional performance and the final output is a probabilistic system performance bound at the task-accomplishment level. The first stage of the Intelligent Hierarchical Decision Architecture is the Clustering Methodology. Here, the raw test data are transformed into a Composite Fuzzy Membership Function--a fuzzy distribution--for further processing. A Composite Fuzzy Membership Function is formed for each test measurement by first, defining Basic Membership Functions that divide the universe of discourse for the variable into relevant fuzzy sets.