on hadamard s maximal determinant problem

On Hadamards Maximal Determinant Problem Judy-anne Osborn MSI, ANU - PowerPoint PPT Presentation

On Hadamards Maximal Determinant Problem Judy-anne Osborn MSI, ANU April 2009 Judy-anne Osborn MSI, ANU On Hadamards Maximal Determinant Problem 0 1 0 1 0 0 0 1 1 . . . . . . . . . . . . . . .


  1. On Hadamard’s Maximal Determinant Problem Judy-anne Osborn MSI, ANU April 2009 Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  2. 0 1 0 1 0 0 0 1   1 . . . . . . . . .   . . . . . . . . .   1   . . . . . . . . .   0 0   . . . . . . . . .   0 m   . . . . . . . . .   1   . . . . . . . . .   0   . . . . . . . . .   1   . . . . . . . . . . . . . . . . . . m max det =? Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  3. A Naive Computer Search Order max det Time 1 1 fast 2 1 fast 3 2 fast 4 3 fast 5 5 fast 6 9 order of days 7 32 order of years 8 56 order of the age of the Universe Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  4. As well by hand? I found nested Max Dets ...   1 0 1 0 0 0 5 } 1 1 0 1 0 0 3 }     0 1 1 0 1 0 2 }   9   0 0 1 1 0 1 } }     1 0 0 1 1 0   1 } 1 1 1 0 0 1 1 Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  5. As well by hand? I found nested Max Dets ...   1 0 1 0 0 0 5 } 1 1 0 1 0 0 3 }     0 1 1 0 1 0 2 }   9   0 0 1 1 0 1 } }     1 0 0 1 1 0   1 } 1 1 1 0 0 1 1 Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  6. As well by hand? I found nested Max Dets ...   1 0 1 0 0 0 5 } 1 1 0 1 0 0 3 }     0 1 1 0 1 0 2 }   9   0 0 1 1 0 1 } }     1 0 0 1 1 0   1 } 1 1 1 0 0 1 1 Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  7. As well by hand? I found nested Max Dets ...   1 0 1 0 0 0 5 } 1 1 0 1 0 0 3 }     0 1 1 0 1 0 2 }   9   0 0 1 1 0 1 } }     1 0 0 1 1 0   1 } 1 1 1 0 0 1 1 Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  8. As well by hand? I found nested Max Dets ...   1 0 1 0 0 0 5 } 1 1 0 1 0 0 3 }     0 1 1 0 1 0 2 }   9   0 0 1 1 0 1 } }     1 0 0 1 1 0   1 } 1 1 1 0 0 1 1 ◮ But no further! Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  9. The problem turns out to be famous ◮ Hadamard’s Maximal Determinant Problem was posed in 1893 Jacques Hadamard Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  10. ◮ A little selected history on this century-old question ... Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  11. Observe: ◮ 2D: � 1 � 1 � 0 � or 1 � 1 � � 1 � 0 0 � 1 � � � 1 0 1 1 0 1 0 Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  12. Geometry: max | det | = max (hyper-)Volume ◮ 3D:   0 1   1     1 1 0 0     1 1   1 1   0   1 0 1 1 1 0   0 1 1 Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  13. An equivalent problem: { +1 , − 1 } matrices ( ( ( ( 1 1 1 1 1 1 1 1 ( ( ( ( 1 0 1 2 0 2 0 2 0 2 1 1 1 1 − − − − − − × ( − 2) border add row 1 1 1 0 2 2 0 0 2 2 0 1 1 1 1 − − − − − − 0 1 1 0 2 2 0 0 2 2 1 1 1 1 − − − − − − m × m matrix column ops ( 1 1 1 1 ( − 1 1 1 1 − − − 1 1 1 1 − 1 1 1 1 − row ops ( ( 1 1 1 1 − 1 1 1 1 − − − 1 1 1 1 − − − 1 1 1 1 − Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  14. An equivalent problem: { +1 , − 1 } matrices ( ( ( ( 1 1 1 1 1 1 1 1 ( ( ( ( 1 0 1 2 0 2 0 2 0 2 1 1 1 1 − − − − − − × ( − 2) border add row 1 1 1 0 2 2 0 0 2 2 0 1 1 1 1 − − − − − − 0 1 1 0 2 2 0 0 2 2 1 1 1 1 − − − − − − m × m matrix column ops ( 1 1 1 1 ( − 1 1 1 1 − − − 1 1 1 1 − 1 1 1 1 − | det new | = 2 m | det old | row ops ( ( 1 1 1 1 − 1 1 1 1 − − − 1 1 1 1 − − − 1 1 1 1 − Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  15. Volume interpretation ⇒ upper bound on | max det | � � } � ± 1 · · · ± 1 � � � � . . ... . . max � � n . . � � � � ± 1 · · · ± 1 � } n ◮ What is the upper bound? Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  16. Volume interpretation ⇒ upper bound on | max det | � � � } ± 1 · · · ± 1 � � � � . . ... . . max � � n . . � � � � ± 1 · · · ± 1 � } n ◮ What is the upper bound? � n �� ( ± 1) 2 + · · · + ( ± 1) 2 = n n / 2 Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  17. Volume interpretation ⇒ upper bound on | max det | � � � } ± 1 · · · ± 1 � � � � . . ... . . max � � n . . � � � � ± 1 · · · ± 1 � } n ◮ What is the upper bound? � n �� ( ± 1) 2 + · · · + ( ± 1) 2 = n n / 2 ◮ Why? Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  18. Volume interpretation ⇒ upper bound on | max det | � � � } ± 1 · · · ± 1 � � � � . . ... . . max � � n . . � � � � ± 1 · · · ± 1 � } n ◮ What is the upper bound? � n �� ( ± 1) 2 + · · · + ( ± 1) 2 = n n / 2 ◮ Why? (Columns/rows orthogonal) Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  19. When is the bound tight? ◮ Tight when { +1 , − 1 } square matrix H of order n satisfies HH T = nI Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  20. When is the bound tight? ◮ Tight when { +1 , − 1 } square matrix H of order n satisfies HH T = nI ◮ H is called a Hadamard Matrix . Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  21. When is the bound tight? ◮ Tight when { +1 , − 1 } square matrix H of order n satisfies HH T = nI ◮ H is called a Hadamard Matrix . ◮ A necessary condition on existence of H is: n = 1 , 2 or n ≡ 0 (mod 4) Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  22. When is the bound tight? ◮ Tight when { +1 , − 1 } square matrix H of order n satisfies HH T = nI ◮ H is called a Hadamard Matrix . ◮ A necessary condition on existence of H is: n = 1 , 2 or n ≡ 0 (mod 4) ◮ Hadamard Conjecture (Paley, 1933): this is also sufficient . Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  23. Evidence for Hadamard Conjecture ◮ Many constructions for infinite families, including ◮ Sylvester, ∀ 2 r ◮ First Paley, using finite fields, ∀ p r + 1, p prime ◮ Second Paley, using finite fields, ∀ 2 p r + 2, p prime Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  24. Evidence for Hadamard Conjecture ◮ Many constructions for infinite families, including ◮ Sylvester, ∀ 2 r ◮ First Paley, using finite fields, ∀ p r + 1, p prime ◮ Second Paley, using finite fields, ∀ 2 p r + 2, p prime ◮ Other ‘constructions’ and ‘ad hoc’ examples due to people including ◮ Williamson ◮ Jenny Seberry Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  25. Evidence for Hadamard Conjecture ◮ Many constructions for infinite families, including ◮ Sylvester, ∀ 2 r ◮ First Paley, using finite fields, ∀ p r + 1, p prime ◮ Second Paley, using finite fields, ∀ 2 p r + 2, p prime ◮ Other ‘constructions’ and ‘ad hoc’ examples due to people including ◮ Williamson ◮ Jenny Seberry ◮ Smallest n ≡ 0 (mod 4) currently undecided: n = 668 . Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  26. More evidence Order Number of inequivalent Hadamard matrices – see Sloan’s sequence A007299 1 1 2 1 4 1 8 1 12 1 16 5 20 3 24 60 28 487 32 ≥ 3 578 006 36 ≥ 18 292 717 Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  27. Max Dets for non-Hadamard orders? Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  28. Max Dets for non-Hadamard orders? n ≡ 1 1 5 9 13 17 21 25 | max det | 3 × 1 1 7 × 2 3 15 × 3 5 20 × 4 7 29 × 5 9 42 × 6 11 1 2 n − 1 n ≡ 2 2 6 10 14 18 | max det | 5 × 1 1 18 × 2 3 39 × 3 5 68 × 4 7 1 2 n − 1 n ≡ 3 3 7 11 15 | max det | 9 × 1 1 40 × 2 3 105 × 3 5 1 2 n − 1 Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

  29. Max Dets for non-Hadamard orders? n ≡ 1 1 5 9 13 17 21 25 | max det | 3 × 1 1 7 × 2 3 15 × 3 5 20 × 4 7 29 × 5 9 42 × 6 11 1 2 n − 1 n ≡ 2 2 6 10 14 18 | max det | 5 × 1 1 18 × 2 3 39 × 3 5 68 × 4 7 1 2 n − 1 n ≡ 3 3 7 11 15 | max det | 9 × 1 1 40 × 2 3 105 × 3 5 1 2 n − 1 Judy-anne Osborn MSI, ANU On Hadamard’s Maximal Determinant Problem

Recommend


More recommend