A Dynamic Clustering Based on Hybrid PS-ACO for Recognizing Oil-Bearing Reservoir

A dynamic clustering algorithm based on hybrid particle swarm-ant colony optimization (PS-ACO) algorithm is presented in the paper. In the algorithm, the number of cluster is dynamic, ACO algorithm is modified by particle swarm optimization (PSO), both the external function and internal function are used to measure the quality evaluation for clustering. The optimal partition is fulfilled by improved PS-ACO algorithm. With its application in recognizing oil-bearing reservoir, the result of simulation indicates that Jaccard index, the external function, is maximum and the internal function, the sum of variance between the object and the center in a cluster is minimum when the cluster number is four. Thus the algorithm has the preferable capability in forecasting and verifying aspects in recognizing oil-bearing reservoir.