A Quorum Sensing Inspired Algorithm for Dynamic Clustering

摘要

Quorum sensing is a decentralized biological process, through which acommunity of cells with no global awareness coordinate their functionalbehaviors based solely on cell-medium interactions and local decisions. Thispaper draws inspirations from quorum sensing and colony competition to derive anew algorithm for data clustering. The algorithm treats each data as a singlecell, and uses knowledge of local connectivity to cluster cells into multiplecolonies simultaneously. It simulates auto-inducers secretion in quorum sensingto tune the influence radius for each cell. At the same time, sparselydistributed core cells spread their influences to form colonies, andinteractions between colonies eventually determine each cell's identity. Thealgorithm has the flexibility to analyze not only static but also time-varyingdata, which surpasses the capacity of many existing algorithms. Its stabilityand convergence properties are established. The algorithm is tested on severalapplications, including both synthetic and real benchmarks data sets, allelesclustering, community detection, image segmentation. In particular, thealgorithm's distinctive capability to deal with time-varying data allows us toexperiment it on novel applications such as robotic swarms grouping andswitching model identification. We believe that the algorithm's promisingperformance would stimulate many more exciting applications.