A multilevel k-way partitioning algorithm for finite element meshes using competing ant colonies

被引:0
作者
Langham, AE [1 ]
Grant, PW [1 ]
机构
[1] Univ Wales, Dept Comp Sci, Swansea SA2 8PP, W Glam, Wales
来源
GECCO-99: PROCEEDINGS OF THE GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE | 1999年
关键词
D O I
暂无
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
The self-organizing properties of ant colonies are employed to tackle the classical combinatorial optimization problem of graph partitioning. Structural information from the graph is mapped onto an environment upon which a number of colonies compete for re sources. Using Genetic Programming a Foraging Strategy is evolved which when executed by the ants in each colony leads to a restructuring of the global environment corresponding to a good partition. Multiple colonies allows for simultaneous k-way partitioning which can provide better partitions than current algorithms which are based on recursive bisection.
引用
收藏
页码:1602 / 1608
页数:7
相关论文
共 11 条
  • [1] [Anonymous], 1993, MULTILEVEL ALGORITHM
  • [2] Grasse P. P., 1959, Insectes Sociaux Paris, V6, P41, DOI 10.1007/BF02223791
  • [3] HENDRICKSON B, 1995, SIAM J SCI COMPUTING, V16
  • [4] Hendrickson Bruce, 1993, Technical Report
  • [5] Kernighan B. W., 1970, Bell System Technical Journal, V49, P291
  • [6] Koza JR, 1992, Genetic programming
  • [7] KUNTZ P, 1994, 3 INT C SIM AD BEH A, V3
  • [8] KUNTZ P, 1997, 4 EUR C ART LIF JUL
  • [9] LASZEWSKI G, 1991, 4 INT C GEN ALG JUL
  • [10] Simon H. D., 1991, Computing Systems in Engineering, V2, P135, DOI 10.1016/0956-0521(91)90014-V