Please make sure you have installed Git Large File Storage when you clone this repository to be able to down the data files.

This is the repository for the Dynamically Optimized Unstructured Grid algorithm. The associated paper is

Weiming Hu, Guido Cervone, Dynamically Optimized Unstructured Grid (DOUG) for Analog Ensemble of Numerical Weather Predictions Using Evolutionary Algorithms.

The repository also provides the source code for the R package RAnEn which is used to generate part of the test data. The original data are too big to be included as a test. Therefore, precomputed data are provided.

Please make sure R is installed with the following packages available:

• sp
• GA
• maps
• gstat
• ncdf4
• rgdal
• fields
• raster
• deldir
• spatstat
• maptools
• stringr
• RColorBrewer
• velox
• RAnEn

To install RAnEn, please run the following command in R:

install.packages("https://github.com/Weiming-Hu/AnalogsEnsemble/raw/master/RAnalogs/releases/RAnEn_latest.tar.gz", repos = NULL)

More information about RAnEn can be found here.

Please make sure you have installed Git Large File Storage when you clone this repository to be able to down the data files.

Please source the R script main.R to run a test case. Parameters can be set in several files separately:

• different-DOUG.R
• GA_core.R
• GA_real-data.R
• main.R

This main.R runs three different algorithms (Pure random, Random + elitism, and DOUG) to compare the performance. The final step of the script create a figure for the optimized unstructured grid.