Numerical comparison of network design algorithms for regionalized variables
The technique of kriging has a fundamental importance in applied sciences such as hydrology, meteorology, soil sciences, and mining. By using kriging, not only can the estimates of the natural phenomena be determined, but the estimation variances reflect the uncertainty of the estimation process. The sampling points for kriging should be selected to minimize the uncertainty, that is, to minimize the estimation variance of the kriging estimator. In this paper four algorithms for optimal observation network design are compared. Two of the algorithms give global optima, while the others give only suboptimal solutions. The computer time required for using the heuristic algorithms, giving only suboptimal solutions, is much less than that of the optimizing procedures. It is also shown on the basis of our experiments that the suboptimal solutions are either optimal or very close to optimal; consequently on the basis of our simulated examples, the heuristic algorithms are highly recommended for practical applications.
- Local optimization of observed points with no need to observations
- Synthesis of structure of regional computer networks under additional conditions
- scientific article; zbMATH DE number 3869085 (Why is no real title available?)
- Comparison of the genetic algorithm and incremental optimisation routines for a Bayesian inverse modelling based network design
- Snow water equivalent estimation using blackbox optimization
- The updated kriging variance and optimal sample design
- On the equivalence of kriging and maximum entropy estimators
- Some aspects of detection networks optimization with the kriging procedure
This page was built for publication: Numerical comparison of network design algorithms for regionalized variables
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1081294)