Average performance of adaptive algorithms for programming co-ordinate measuring machines under budget constraint

In this paper we develop two adaptive algorithms for programming co-ordinate measuring machines assuming fixed sampling budget. Two different costs are considered: the travelling cost of the machine probe, and the sampling cost to read and store all measurements. Simulation is used to compare the average performance of the proposed algorithms under the assumption of Wiener measure on the space of all surface contours of the manufactured parts. Expected value of the probability of Type II error is the criterion that we use to characterize algorithms performance. Analysis shows that placing sample points according to the criterion of maximizing the expected gain demonstrates a substantial improvement in the average performance.