Knowledge evolutionary algorithm based on granular computing

Granular computing makes mainly use of the information of different granularities and hierarchies to solve problems of the uncertain, fuzzy, imprecise, part true and a number of information. This paper has analyzed the evolutionary characteristics of knowledge granulation and has proposed the evolution algorithm of knowledge granulation (EAKG). EAKG algorithm applies knowledge granulation to genetic programming and carries through the evaluation according to coverage degree and depends on degree to obtain some new rules. In addition, this paper has also given the recursive model of knowledge granulation evolution, crossover operator and mutation operator, etc. Through the experiments it has proved that it is the reasonable and effective to carry out solution of knowledge evolution with granule computing.

[1]  Guillermo Leguizamón,et al.  Evolution of classification rules for comprehensible knowledge discovery , 2007, 2007 IEEE Congress on Evolutionary Computation.

[2]  Yi Pan,et al.  FIK Model: Novel Efficient Granular Computing Model for Protein Sequence Motifs and Structure Information Discovery , 2006, Sixth IEEE Symposium on BioInformatics and BioEngineering (BIBE'06).

[3]  Xue Zhen-gui.  A Knowledge-Based Hierachical Genetic Algorithm , 2006 .

[4]  Yi Pan,et al.  Novel efficient granular computing models for protein sequence motifs and structure information discovery , 2009, Int. J. Comput. Biol. Drug Des..

[5]  Emiliano Carre.  Evolution of Classification Rules for Comprehensible Knowledge Discovery , 2007 .

[6]  A. Bourmistrova,et al.  Control system design optimisation via genetic programming , 2007, 2007 IEEE Congress on Evolutionary Computation.

[7]  Jingtao Yao,et al.  Information granulation and granular relationships , 2005, 2005 IEEE International Conference on Granular Computing.

[8]  Guoyin Wang,et al.  A rule generation algorithm based on granular computing , 2005, 2005 IEEE International Conference on Granular Computing.