Title :
Divide and conquer in granular computing topological partitions
Author_Institution :
Dept. of Comput. Sci., San Jose State Univ., CA, USA
Abstract :
Divide and conquer are classical problem solving strategy. Implicitly the divide is a mathematical partitioning; no overlapping is allowed. However, in practices the divide cannot be very clean; some degree of overlapping is unavoidable. A binary granulation, which is a granulation defined by a binary relation, is not a partition; so overlapping does exist. However, by looking at the problem skillfully a binary granulation can be interpreted as a topological partition. By a topological partition is a partition in which each equivalence class has a neighborhood. These neighborhoods do overlap. So a granulation is a "topological divide and conquer".
Keywords :
divide and conquer methods; equivalence classes; topology; binary granulation; divide and conquer; equivalence class; granular computing; mathematical partitioning; problem solving strategy; topological partition; Computer science; History; Humans; Information retrieval; Magnetic heads; Neck; Problem-solving; Relational databases; Rough sets; Topology;
Conference_Titel :
Fuzzy Information Processing Society, 2005. NAFIPS 2005. Annual Meeting of the North American
Print_ISBN :
0-7803-9187-X
DOI :
10.1109/NAFIPS.2005.1548548