### Abstract

Many fundamental problems in scientific computing have more than one solution method. It is not uncommon for alternative solution methods to represent different tradeoffs between solution cost and reliability. Furthermore, the performance of a solution method often depends on the numerical properties of the problem instance and thus can vary dramatically across application domains. In such situations, it is natural to consider the construction of a multi-method composite solver to potentially improve both the average performance and reliability. In this paper, we provide a combinatorial framework for developing such composite solvers. We provide analytical results for obtaining an optimal composite from a set of methods with normalized measures of performance and reliability. Our empirical results demonstrate the effectiveness of such optimal composites for solving large, sparse linear systems of equations.

Original language | English (US) |
---|---|

Title of host publication | Computational Science, ICCS 2002 - International Conference, Proceedings |

Pages | 325-334 |

Number of pages | 10 |

Edition | PART 2 |

State | Published - Dec 1 2002 |

Event | International Conference on Computational Science, ICCS 2002 - Amsterdam, Netherlands Duration: Apr 21 2002 → Apr 24 2002 |

### Publication series

Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|

Number | PART 2 |

Volume | 2330 LNCS |

ISSN (Print) | 0302-9743 |

ISSN (Electronic) | 1611-3349 |

### Conference

Conference | International Conference on Computational Science, ICCS 2002 |
---|---|

Country | Netherlands |

City | Amsterdam |

Period | 4/21/02 → 4/24/02 |

### Fingerprint

### ASJC Scopus subject areas

- Theoretical Computer Science
- Computer Science(all)

### Cite this

*Computational Science, ICCS 2002 - International Conference, Proceedings*(PART 2 ed., pp. 325-334). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2330 LNCS, No. PART 2).

**A combinatorial scheme for developing efficient composite solvers.** / Bhowmick, Sanjukta; Raghavan, Padma; Teranishi, Keita.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

*Computational Science, ICCS 2002 - International Conference, Proceedings.*PART 2 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 2330 LNCS, pp. 325-334, International Conference on Computational Science, ICCS 2002, Amsterdam, Netherlands, 4/21/02.

}

TY - GEN

T1 - A combinatorial scheme for developing efficient composite solvers

AU - Bhowmick, Sanjukta

AU - Raghavan, Padma

AU - Teranishi, Keita

PY - 2002/12/1

Y1 - 2002/12/1

N2 - Many fundamental problems in scientific computing have more than one solution method. It is not uncommon for alternative solution methods to represent different tradeoffs between solution cost and reliability. Furthermore, the performance of a solution method often depends on the numerical properties of the problem instance and thus can vary dramatically across application domains. In such situations, it is natural to consider the construction of a multi-method composite solver to potentially improve both the average performance and reliability. In this paper, we provide a combinatorial framework for developing such composite solvers. We provide analytical results for obtaining an optimal composite from a set of methods with normalized measures of performance and reliability. Our empirical results demonstrate the effectiveness of such optimal composites for solving large, sparse linear systems of equations.

AB - Many fundamental problems in scientific computing have more than one solution method. It is not uncommon for alternative solution methods to represent different tradeoffs between solution cost and reliability. Furthermore, the performance of a solution method often depends on the numerical properties of the problem instance and thus can vary dramatically across application domains. In such situations, it is natural to consider the construction of a multi-method composite solver to potentially improve both the average performance and reliability. In this paper, we provide a combinatorial framework for developing such composite solvers. We provide analytical results for obtaining an optimal composite from a set of methods with normalized measures of performance and reliability. Our empirical results demonstrate the effectiveness of such optimal composites for solving large, sparse linear systems of equations.

UR - http://www.scopus.com/inward/record.url?scp=23044532053&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=23044532053&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:23044532053

SN - 354043593X

SN - 9783540435938

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 325

EP - 334

BT - Computational Science, ICCS 2002 - International Conference, Proceedings

ER -