Seismic Attributes Clustering based on Artificial Bee Colony
Chengyin Cao, Zheng Xiaodong, Jinsong Li · 2014
Seismic attributes clustering is an important approach for seismic attributes analysis in reservoir prediction. K-means is one of the most popular method used in clustering analysis, but it is highly influenced by the initial centroids, and always converged to local optimum solution. Artificial bee colony is one of swarm intelligence methods, and proved to be converged to global optimum, was brought into seismic attributes clustering for optimization the cluster centroids. Through application of field data, the clustering method based on artificial bee colony could avoids the problem of unstable and local optimization, especially could provide more reasonable geological information about the carbonate reservoir feature, such as reef and shoal complex, fluids and fault, etc..