2016 Fiscal Year Annual Research Report
Analyzing the limits of computation using large scale linear programming
Project Area | A multifaceted approach toward understanding the limitations of computation |
Project/Area Number |
24106006
|
Research Institution | Kyoto University |
Principal Investigator |
Avis David 京都大学, 情報学研究科, 研究員 (90584110)
|
Co-Investigator(Kenkyū-buntansha) |
天野 一幸 群馬大学, 大学院理工学府, 教授 (30282031)
|
Project Period (FY) |
2012-06-28 – 2017-03-31
|
Keywords | 幾何計算 / 最適化数 / 数理計画法への応用 |
Outline of Annual Research Achievements |
International Collaboration: We received visits from David Bremner(UNB, Frederiction) from December12-31 and Hans Tiwary (Charles) from February 1-14 at Kyoto University. Bremner and Avis released their compiler for computing linear programs from pseudocode called mts. A working prototype was built and applied to the matching problem. Tiwary and Avis completed a paper giving a compact formulation for 2-SAT even though it has exponential extension complexity.We published it on the arXiv and submitted it to a journal. Both visitors gave invited seminars. Three journal papers were published by Avis during the last year and one new paper was written with Charles Jordan (Hokkaido U.) and uploaded to the arXiv. It presented a description of the mts package and numerical results for parallelizing reverse search algorithms for generating topological sorts and spanning trees. Using data sharing we developed parallel implementations of existing implementations of algorithms for satisfiability testing.This code was released in January 2017. We also made further improvements to mplrs during the year resulting in a new release in November 2016. Luc Devroye (McGill)visited Kyoto from May 25-June 2. We did joint research on analyzing parallel tree search with bucketing on Galton-Watson trees. We published a paper on this on the arXiv in March 2017 and submitted it to a journal. Amano enumerated all NPN equivalence classes of non-degenerate n-variable Boolean functions of sensitivity three and obtained the maximum of their block, certificate, decision tree and degree complexity.
|
Research Progress Status |
28年度が最終年度であるため、記入しない。
|
Strategy for Future Research Activity |
28年度が最終年度であるため、記入しない。
|