ON DISCRETE HESSIAN MATRIX AND CONVEX EXTENSIBILITY

• Moriguchi Satoko

  Advanced Institute of Industrial Technology

• Murota Kazuo

  University of Tokyo

抄録

For functions defined on integer lattice points, discrete versions of the Hessian matrix have been considered in various contexts. In discrete convex analysis, for example certain combinatorial properties of the discrete Hessian matrices are known to characterize M^〓-convex and L^〓-convex functions, which can be extended to convex functions in real variables. The relationship between convex extensibility and discrete Hessian matrices is not fully understood in general, and unfortunately, some vague or imprecise statements have been made in the literature. This note points out that the positive semidefiniteness of the discrete Hessian matrix does not imply nor is implied by convex extensibility of discrete functions.

収録刊行物

• 日本オペレーションズ・リサーチ学会論文誌

  日本オペレーションズ・リサーチ学会論文誌 55 (1), 48-62, 2012

  公益社団法人 日本オペレーションズ・リサーチ学会

キーワード

• Discrete optimization
• discrete convex function
• Hessian matrix
• semidefinite programming

詳細情報

• CRID: 1390001204109876480
• NII論文ID: 110009436526
• NII書誌ID: AA00703935
• DOI: 10.15807/jorsj.55.48
• ISSN: 21888299
• ISSN: 04534514
• NDL書誌ID: 023545332
• Web Site: https://ndlsearch.ndl.go.jp/books/R000000004-I023545332
• Web Site: https://www.jstage.jst.go.jp/article/jorsj/55/1/55_KJ00008017544/_pdf
• 本文言語コード: en
• データソース種別:
  • JaLC
  • NDL
  • Crossref
  • CiNii Articles
• 抄録ライセンスフラグ: 使用不可