Randomized greedy algorithms for the hypergraph partitioning problem

We propose a series of randomized greedy construction schemes for the hypergraph partitioning problem. While the nal results are inferior to those obtained by recent multi-level methods, the advantages of our greedy schemes are their simplicity and low computational complexity. The best greedy algorithms considered obtain low cut values and large standard deviations of the results. Therefore, when independent repetitions are considered, the quality of the best solution greatly improves and, in some cases, it is superior to the variable-depth Fiduccia-Mattheyses (FM) algorithm, for smaller CPU times. Furthermore, the algorithms can be used as building blocks in more complex schemes. For example, we successfully employ our greedy schemes to produce initial partitions for improvement-based heuristics. In particular, if FM is run starting from partitions generated by our greedy schemes, instead of random initial solutions, a signi cant improvement of the average solution quality is achieved for comparable computation times.

[1]  Brian W. Kernighan,et al.  An efficient heuristic procedure for partitioning graphs , 1970, Bell Syst. Tech. J..

[2]  Charles M. Fiduccia,et al.  A linear-time heuristic for improving network partitions , 1988, 25 years of DAC.

[3]  Balakrishnan Krishnamurthy,et al.  An Improved Min-Cut Algonthm for Partitioning VLSI Networks , 1984, IEEE Transactions on Computers.

[4]  Thomas Lengauer,et al.  Combinatorial algorithms for integrated circuit layout , 1990, Applicable theory in computer science.

[5]  Fillia Makedon,et al.  Circuit partitioning into small sets: a tool to support testing with further applications , 1991, Proceedings of the European Conference on Design Automation..

[6]  Jaeseok Kim,et al.  An Efficient Method of Partitioning Circuits for Multiple-FPGA Implementation. , 1993, 30th ACM/IEEE Design Automation Conference.

[7]  Krzysztof Kozminski,et al.  Cost Minimization of Partitions into Multiple Devices , 1993, 30th ACM/IEEE Design Automation Conference.

[8]  Baldomir Zajc,et al.  A unified cost model for min-cut partitioning with replication applied to optimization of large heterogeneous FPGA partitions , 1994, EURO-DAC '94.

[9]  Manuel Laguna,et al.  A Greedy Randomized Adaptive Search Procedure for the Two-Partition Problem , 1994, Oper. Res..

[10]  Baldomir Zajc,et al.  Multi-way Netlist Partitioning into Heterogeneous FPGAs and Minimization of Total Device Cost and Interconnect , 1994, 31st Design Automation Conference.

[11]  Chung-Kuan Cheng,et al.  Circuit Partitioning for Huge Logic Emulation Systems , 1994, 31st Design Automation Conference.

[12]  Kurt Mehlhorn,et al.  LEDA: a platform for combinatorial and geometric computing , 1997, CACM.

[13]  Andrew B. Kahng,et al.  On implementation choices for iterative improvement partitioning algorithms , 1995, EURO-DAC '95/EURO-VHDL '95.

[14]  Andrew B. Kahng,et al.  Recent directions in netlist partitioning: a survey , 1995, Integr..

[15]  Shashi Shekhar,et al.  Partitioning Similarity Graphs: A Framework for Declustering Problems , 1996, Inf. Syst..

[16]  Jaideep Srivastava,et al.  Web Mining: Pattern Discovery from World Wide Web Transactions , 1996 .

[17]  S. Dutt,et al.  A probability-based approach to VLSI circuit partitioning , 1996, 33rd Design Automation Conference Proceedings, 1996.

[18]  S. Dutt,et al.  VLSI circuit partitioning by cluster-removal using iterative improvement techniques , 1996, Proceedings of International Conference on Computer Aided Design.

[19]  Shashi Shekhar,et al.  Multilevel hypergraph partitioning: applications in VLSI domain , 1999, IEEE Trans. Very Large Scale Integr. Syst..