Intelligent scheduling optimization using a rule-based artificial neural network

Requirements for greater precision and reduced rejection/acceptance errors demand improved inspection methods that can be provided by implementing increased automation into the quality control (QC) process. Automated inspection planners can require significant human interface work deciphering the generated output, often requiring more work than the service they provide. The output of an automated inspection planner needs to produce a process plan that would resemble one created by an expert inspector. This paper discusses the implementation of an artificial neural network (ANN) to optimize the sequence of inspection points based upon inspection rule criteria. Using features from a feature-based CAD, the automated inspection planner uses the inspection rules to create "rule matrices" that the ANN will use to sequence the inspection points to be probed by a coordinate measurement machine (CMM). This is an extension of the familiar traveling salesman problem (TSP) solved by a Hopfield neural network.<>