Ant colony ATTA clustering algorithm of rock mass structural plane in groups

Abstract

Based on structural surface normal vector spherical distance and the pole stereographic projection Euclidean distance, two distance functions were established. The cluster analysis of structure surface was conducted by the use of ATTA clustering methods based on ant colony piles, and Silhouette index was introduced to evaluate the clustering effect. The clustering analysis of the measured data of Sanshandao Gold Mine shows that ant colony ATTA-based clustering method does better than K-mean clustering analysis. Meanwhile, clustering results of ATTA method based on pole Euclidean distance and ATTA method based on normal vector spherical distance have a great consistence. The clustering results are most close to the pole isopycnic graph. It can efficiently realize grouping of structural plane and determination of the dominant structural surface direction. It is made up for the defects of subjectivity and inaccuracy in icon measurement approach and has great engineering value.

Keywords

rock mass discontinuity / cluster analysis / ant colony ATTA algorithm / distance function / Silhouette index

References

[1] WangW-xingRock mechanics [M], 2004, Changsha, China, Central South University Press: 52-68

[2] DorigoM, ColonyA. Guest editorial: Special section on ant colony optimization [J]. IEEE Transactions on Evolutionary Computation, 2002, 6(4): 317-319

[3] HandlJ, MeyerB. Improved ant-based clustering and sorting in a document retrieval interface [C]. Proceedings of the Seventh International Conference on Parallel Problem Solving from Nature (PPSN VII), 2002, Berlin, Germany, Springer-Verlag: 913-923

[4] FanL, WangL-q, TangH-ming. Dynamic cluster analysis of discontinuity orientations of jointed rock mass [J]. Rock and Soil Mechanics, 2007, 28(11): 2405-2408

[5] TengJ-d, XuG-l, ShenY-jun. Fuzzy c-means cluster analysis based on ant colony algorithm for orientations of joint [J]. Safety and Environmental Engineering, 2010, 17(4): 114-117

[6] PengA, ChengZ-bin. Study on grouping of structural plane of rock mass based on artificial neural network [J]. Yangtze River, 2011, 42(22): 96-100

[7] JohnsonS C. Hierarchical clustering schemes [J]. Psychometrika, 1967, 32(3): 241-254

[8] DasguptaS. Performance guarantees for hierarchical clustering [M]. Computational Learning Theory, 2002351-363

[9] DeneubourgJ L, GossS, FranksN. The dynamics of collective sorting robot-like ants and ant-like robots [C]. Proceedings of the First International Conference on Simulation of Adaptive Behavior on From Animals to Animals, 1991, Cambridge MA, MIT Press: 356-363

[10] LumerE, FaietaB. Diversity and adaptation in populations of clustering ants [C]. Proceedings of the third International Conference on Simulation of Adaptive Behavior: From Animals to Animals 3, 1994, Cambridge, MA, MIT Press: 501-508

[11] HandlJ, KnowlesJ, DorigoM. On the performance of ant-based clustering [C]. Proc of the 3rd Int Conf on Hybrid Intelligent Systems, 2003, Australia, IOS Press: 204-213

[12] TanaS C, TingK M, TengS W. Simplifying and improving ant-based clustering [J]. Procedia Computer Science, 2011, 4: 46-55

[13] HandlJ K J. Ant-based clustering and topographic mapping [J]. Artificial Life, 2006, 12(1): 35-61

[14] WangJ-b, TuA-l, HuangH-wei. An ant colony clustering algorithm improved from ATTA [J]. Physics Procedia, 2012, 24: 1414-1421

[15] RousseeuwP J. Silhouettes: A graphical aid to the interpretation and validation of cluster analysis [J]. Comp App Math, 1987, 20: 53-65

[16] HammahR E, CurranJ H. Fuzzy cluster algorithm for the automatic identification of joint sets [J]. Int J Rock Mech Min Sci, 1998, 35(7): 889-905