new upper bounds for nonbinary codes based on
play

New upper bounds for nonbinary codes based on semidefinite - PowerPoint PPT Presentation

New upper bounds for nonbinary codes based on semidefinite programming and parity Sven Polak Partly based on joint work with Bart Litjens and Lex Schrijver Korteweg-de Vries Institute for Mathematics Faculty of Science University of Amsterdam


  1. New upper bounds for nonbinary codes based on semidefinite programming and parity Sven Polak Partly based on joint work with Bart Litjens and Lex Schrijver Korteweg-de Vries Institute for Mathematics Faculty of Science University of Amsterdam Plzeˇ n, October 6th, 2016 Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  2. Definitions and notation Fix q , n , d ∈ N with q ≥ 2. Define [ q ] := { 0 , . . . , q − 1 } . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  3. Definitions and notation Fix q , n , d ∈ N with q ≥ 2. Define [ q ] := { 0 , . . . , q − 1 } . A word is an element of [ q ] n and a code is a subset of [ q ] n . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  4. Definitions and notation Fix q , n , d ∈ N with q ≥ 2. Define [ q ] := { 0 , . . . , q − 1 } . A word is an element of [ q ] n and a code is a subset of [ q ] n . The Hamming distance between two words u , v ∈ [ q ] n is d H ( u , v ) := |{ i : u i � = v i }| . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  5. Definitions and notation Fix q , n , d ∈ N with q ≥ 2. Define [ q ] := { 0 , . . . , q − 1 } . A word is an element of [ q ] n and a code is a subset of [ q ] n . The Hamming distance between two words u , v ∈ [ q ] n is d H ( u , v ) := |{ i : u i � = v i }| . The minimum distance d min ( C ) of a code C ⊆ [ q ] n is the minimum of d H ( u , v ) over all distinct u , v ∈ C . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  6. Definitions and notation Fix q , n , d ∈ N with q ≥ 2. Define [ q ] := { 0 , . . . , q − 1 } . A word is an element of [ q ] n and a code is a subset of [ q ] n . The Hamming distance between two words u , v ∈ [ q ] n is d H ( u , v ) := |{ i : u i � = v i }| . The minimum distance d min ( C ) of a code C ⊆ [ q ] n is the minimum of d H ( u , v ) over all distinct u , v ∈ C . Example (i) d min ( { 1112 , 2111 , 3134 } ) = 2, Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  7. Definitions and notation Fix q , n , d ∈ N with q ≥ 2. Define [ q ] := { 0 , . . . , q − 1 } . A word is an element of [ q ] n and a code is a subset of [ q ] n . The Hamming distance between two words u , v ∈ [ q ] n is d H ( u , v ) := |{ i : u i � = v i }| . The minimum distance d min ( C ) of a code C ⊆ [ q ] n is the minimum of d H ( u , v ) over all distinct u , v ∈ C . Example (i) d min ( { 1112 , 2111 , 3134 } ) = 2, Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  8. The parameter A q ( n , d ) Definition A q ( n , d ) := max {| C | | C ⊆ [ q ] n , d min ( C ) ≥ d } . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  9. The parameter A q ( n , d ) Definition A q ( n , d ) := max {| C | | C ⊆ [ q ] n , d min ( C ) ≥ d } . Examples A q ( n , 1) = q n . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  10. The parameter A q ( n , d ) Definition A q ( n , d ) := max {| C | | C ⊆ [ q ] n , d min ( C ) ≥ d } . Examples A q ( n , 1) = q n . A 5 (7 , 6) = 15. Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  11. The parameter A q ( n , d ) Definition A q ( n , d ) := max {| C | | C ⊆ [ q ] n , d min ( C ) ≥ d } . Examples A q ( n , 1) = q n . A 5 (7 , 6) = 15. (i) Tables with bounds on A q ( n , d ) on the website of Andries Brouwer. Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  12. The parameter A q ( n , d ) Definition A q ( n , d ) := max {| C | | C ⊆ [ q ] n , d min ( C ) ≥ d } . Examples A q ( n , 1) = q n . A 5 (7 , 6) = 15. (i) Tables with bounds on A q ( n , d ) on the website of Andries Brouwer. (ii) Interesting parameter in cryptography: a code C ⊆ [ q ] n with d min ( C ) = 2 e + 1 is e - error correcting . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  13. The parameter A q ( n , d ) – II Definition A q ( n , d ) := max {| C | | C ⊆ [ q ] n , d min ( C ) ≥ d } . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  14. The parameter A q ( n , d ) – II Definition A q ( n , d ) := max {| C | | C ⊆ [ q ] n , d min ( C ) ≥ d } . Remark Let G = ( V , E ) be the graph with V = [ q ] n and E := {{ u , v } | 0 < d H ( u , v ) < d } . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  15. The parameter A q ( n , d ) – II Definition A q ( n , d ) := max {| C | | C ⊆ [ q ] n , d min ( C ) ≥ d } . Remark Let G = ( V , E ) be the graph with V = [ q ] n and E := {{ u , v } | 0 < d H ( u , v ) < d } . Then A q ( n , d ) = α ( G ), the stable set number of G . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  16. The parameter A q ( n , d ) – II Definition A q ( n , d ) := max {| C | | C ⊆ [ q ] n , d min ( C ) ≥ d } . Remark Let G = ( V , E ) be the graph with V = [ q ] n and E := {{ u , v } | 0 < d H ( u , v ) < d } . Then A q ( n , d ) = α ( G ), the stable set number of G . 000 001 101 010 n = 3 , d = 2 110 011 111 100 Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  17. The parameter A q ( n , d ) – II Definition A q ( n , d ) := max {| C | | C ⊆ [ q ] n , d min ( C ) ≥ d } . Remark Let G = ( V , E ) be the graph with V = [ q ] n and E := {{ u , v } | 0 < d H ( u , v ) < d } . Then A q ( n , d ) = α ( G ), the stable set number of G . 000 001 101 010 n = 3 , d = 2 110 011 111 100 Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  18. The parameter A q ( n , d ) – II Definition A q ( n , d ) := max {| C | | C ⊆ [ q ] n , d min ( C ) ≥ d } . Remark Let G = ( V , E ) be the graph with V = [ q ] n and E := {{ u , v } | 0 < d H ( u , v ) < d } . Then A q ( n , d ) = α ( G ), the stable set number of G . 000 001 101 010 n = 3 , d = 2 110 011 111 100 Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  19. The parameter A q ( n , d ) – II Definition A q ( n , d ) := max {| C | | C ⊆ [ q ] n , d min ( C ) ≥ d } . Remark Let G = ( V , E ) be the graph with V = [ q ] n and E := {{ u , v } | 0 < d H ( u , v ) < d } . Then A q ( n , d ) = α ( G ), the stable set number of G . 000 001 101 010 n = 3 , d = 2 110 011 111 100 Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  20. Delsarte bound Notation Let C k be the collection of codes C ⊆ [ q ] n with | C | ≤ k . Given x : C 2 → R ≥ 0 , define the C 1 × C 1 -matrix M x by ( M x ) C , C ′ = x ( C ∪ C ′ ) . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  21. Delsarte bound Notation Let C k be the collection of codes C ⊆ [ q ] n with | C | ≤ k . Given x : C 2 → R ≥ 0 , define the C 1 × C 1 -matrix M x by ( M x ) C , C ′ = x ( C ∪ C ′ ) . It can be proven that the Delsarte bound equals � � θ q ( n , d ) := max x ( { v } ) | x : C 2 → R ≥ 0 with: v ∈ [ q ] n (i) x ( ∅ ) = 1 , (ii) x ( C ) = 0 if d min ( C ) < d , � (iii) M x is positive semidefinite . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  22. The quadruple bound Notation Let C k be the collection of codes C ⊆ [ q ] n with | C | ≤ k . Given x : C 4 → R ≥ 0 , define the C 2 × C 2 -matrix M x by ( M x ) C , C ′ = x ( C ∪ C ′ ) . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  23. The quadruple bound Notation Let C k be the collection of codes C ⊆ [ q ] n with | C | ≤ k . Given x : C 4 → R ≥ 0 , define the C 2 × C 2 -matrix M x by ( M x ) C , C ′ = x ( C ∪ C ′ ) . Now we define Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  24. The quadruple bound Notation Let C k be the collection of codes C ⊆ [ q ] n with | C | ≤ k . Given x : C 4 → R ≥ 0 , define the C 2 × C 2 -matrix M x by ( M x ) C , C ′ = x ( C ∪ C ′ ) . Now we define � � x ( { v } ) | x : C 4 → R ≥ 0 with: B q ( n , d ) := max v ∈ [ q ] n (i) x ( ∅ ) = 1 , (ii) x ( C ) = 0 if d min ( C ) < d , � (iii) M x is positive semidefinite . Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  25. The quadruple bound � � x ( { v } ) | x : C 4 → R ≥ 0 with: B q ( n , d ) := max v ∈ [ q ] n (i) x ( ∅ ) = 1 , (ii) x ( C ) = 0 if d min ( C ) < d , � (iii) M x is positive semidefinite . Proposition. A q ( n , d ) ≤ B q ( n , d ) Proof. Let C ⊆ [ q ] n be a code of minimum distance at least d Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  26. The quadruple bound � � x ( { v } ) | x : C 4 → R ≥ 0 with: B q ( n , d ) := max v ∈ [ q ] n (i) x ( ∅ ) = 1 , (ii) x ( C ) = 0 if d min ( C ) < d , � (iii) M x is positive semidefinite . Proposition. A q ( n , d ) ≤ B q ( n , d ) Proof. Let C ⊆ [ q ] n be a code of minimum distance at least d and maximum size. Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  27. The quadruple bound � � x ( { v } ) | x : C 4 → R ≥ 0 with: B q ( n , d ) := max v ∈ [ q ] n (i) x ( ∅ ) = 1 , (ii) x ( C ) = 0 if d min ( C ) < d , � (iii) M x is positive semidefinite . Proposition. A q ( n , d ) ≤ B q ( n , d ) Proof. Let C ⊆ [ q ] n be a code of minimum distance at least d and maximum size. Define x by x ( S ) = 1 if S ⊆ C and x ( S ) = 0 else. Sven Polak Code bounds Plzeˇ n, October 6th, 2016

  28. The quadruple bound � � x ( { v } ) | x : C 4 → R ≥ 0 with: B q ( n , d ) := max v ∈ [ q ] n (i) x ( ∅ ) = 1 , (ii) x ( C ) = 0 if d min ( C ) < d , � (iii) M x is positive semidefinite . Proposition. A q ( n , d ) ≤ B q ( n , d ) Proof. Let C ⊆ [ q ] n be a code of minimum distance at least d and maximum size. Define x by x ( S ) = 1 if S ⊆ C and x ( S ) = 0 else. Then x is feasible. Sven Polak Code bounds Plzeˇ n, October 6th, 2016

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend