IECM 12.0 beta User Manual > Using the IECM > Analysis Tools > Uncertainty > Configure > Sampling Method > Hammersley |
|
The Hammersley sequence sampling technique is more efficient than either the Monte Carlo or Latin-Hypercube sampling techniques. The sampling method is loosely based on the Monte Carlo method. However, instead of using a random number generator, it uses a quasi-random number generator based on Hammersley points to uniformly sample a unit hypercube. These points are an optimal design for placing n points on a k-dimensional hypercube. The sample points are then inverted over a cumulative probability distribution to define the sample set for any uncertainty variable.
Hammersley has the advantage of high precision and consistent behavior in addition to better computational efficiency. The method reduces the number of samples required relative to the other sampling methods for calculating uncertainty by a factor of 2 to 100. The actual sample reduction varies with the uncertainty function being sampled.
See also: |
Diwekar, U.M. and J.R. Kalagnanam, (1997) "Efficient Sampling Technique for Optimization under Uncertainty," AIChE Journal, Vol. 43, No. 2, pp. 440-7. |
Copyright © 2023-2024 University of Wyoming. All rights reserved. Visit us at https://www.iecm-online.com/