Information-based inspection allocation for real-time inspection systems

The use of real-time Automated Visual Inspection (AVI) in the electronics assembly industry is subject to cycle time constraints determined by the manufacturing system. These constraints must be met to maintain the nominal production rate. A problem thus created is determining which components to inspect at each one of the AVI stations such that the inspection time available is used in an optimal way. This paper presents a real-time inspection allocation that is based on the information gained by inspecting one additional component. The selection of which components to inspect is modeled as an information maximization problem. A modified Knapsack greedy heuristic is used to find near-optimal solutions to this optimization problem within the required time constraints.