improvement of the sunflower bound for 1 intersecting
play

Improvement of the sunflower bound for 1-intersecting constant - PowerPoint PPT Presentation

Improvement of the sunflower bound for 1-intersecting constant dimension subspace codes Leo Storme Ghent University Dept. of Mathematics Krijgslaan 281 9000 Ghent Belgium (joint work with D. Bartoli, A. Riet and P . Vandendriessche)


  1. Improvement of the sunflower bound for 1-intersecting constant dimension subspace codes Leo Storme Ghent University Dept. of Mathematics Krijgslaan 281 9000 Ghent Belgium (joint work with D. Bartoli, A. Riet and P . Vandendriessche) Istanbul, November 6, 2015 Leo Storme Constant dimension subspace codes

  2. t -I NTERSECTING CONSTANT DIMENSION SUBSPACE CODES t -Intersecting constant dimension subspace code: Codewords are k -dimensional vector spaces. Distinct codewords intersect in t -dimensional vector spaces. Classical example: Sunflower: all codewords pass through same t -dimensional vector space. Leo Storme Constant dimension subspace codes

  3. S UNFLOWER Leo Storme Constant dimension subspace codes

  4. L ARGE t - INTERSECTING CONSTANT DIMENSION SUBSPACE CODES T HEOREM Large t-intersecting constant dimension subspace codes are sunflowers. If � q k − q t � 2 � q k − q t � | C | > + + 1 , q − 1 q − 1 then C is sunflower. Leo Storme Constant dimension subspace codes

  5. C ONJECTURE Conjecture: Let C be t -intersecting constant dimension subspace code. If | C | > q k + q k − 1 + · · · + q + 1 , then C is sunflower. Leo Storme Constant dimension subspace codes

  6. C OUNTEREXAMPLES TO CONJECTURE Code C of 1-intersecting 3-dimensional spaces in V ( 6 , 2 ) . Conjecture: If | C | > 15, then C is sunflower. Counterexample 1: (Etzion and Raviv) Code C of size 16 which is not sunflower. Counterexample 2: (Bartoli and Pavese) Code C of 1-intersecting 3-dimensional spaces in V ( 6 , 2 ) has size at most 20, and unique example of size 20. Leo Storme Constant dimension subspace codes

  7. I MPROVEMENT TO UPPER BOUND FOR t = 1 If � 2 � q k − q t � q k − q t � | C | > + + 1 , q − 1 q − 1 then C is sunflower. For t = 1, if � q k − q � 2 � q k − q � | C | > + + 1 , q − 1 q − 1 then C is sunflower. Question: Can this bound be improved? Leo Storme Constant dimension subspace codes

  8. I MPROVEMENT TO UPPER BOUND FOR t = 1 Assumptions: C = 1-intersecting constant dimension code of k -spaces. C not sunflower. � q k − q � 2 � q k − q � | C | = + + 1 − δ, q − 1 q − 1 with δ = q k − 2 . Leo Storme Constant dimension subspace codes

  9. I MPROVEMENT TO UPPER BOUND FOR t = 1 See codeword c ∈ C as PG ( k − 1 , q ) . Define S = ∪ c ∈ C c . L EMMA Point P ∈ S belongs to at most q k − 1 q − 1 codewords. Leo Storme Constant dimension subspace codes

  10. I MPROVEMENT TO UPPER BOUND FOR t = 1 Leo Storme Constant dimension subspace codes

  11. I MPROVEMENT TO UPPER BOUND FOR t = 1 L EMMA If | C | > ( q k − q q − 1 ) 2 , then every codeword in C has at least one point in q k − 1 q − 1 codewords. L EMMA If point P lies in q k − 1 q − 1 codewords, then line through P and other point of S is completely contained in S . Leo Storme Constant dimension subspace codes

  12. I MPROVEMENT TO UPPER BOUND FOR t = 1 Leo Storme Constant dimension subspace codes

  13. I MPROVEMENT TO UPPER BOUND FOR t = 1 L EMMA If point P lies in q k − 1 q − 1 codewords, then � q k − q � 2 � q k − q � |S| = | ∪ c ∈ C c | = + + 1 . q − 1 q − 1 REMARK: � q k − q � 2 � q k − q � + + 1 � = | PG ( T , q ) | . q − 1 q − 1 | PG ( 2 k − 2 , q ) | < |S| < | PG ( 2 k − 1 , q ) | . Leo Storme Constant dimension subspace codes

  14. I MPROVEMENT TO UPPER BOUND FOR t = 1 L EMMA If more than q 2 k − 3 + q 2 k − 4 + · · · + q + 1 points of S lie in q k − 1 q − 1 codewords, then ( 2 k − 2 ) -dimensional subspace contained in S . Leo Storme Constant dimension subspace codes

  15. I MPROVEMENT TO UPPER BOUND FOR t = 1 Leo Storme Constant dimension subspace codes

  16. I MPROVEMENT TO UPPER BOUND FOR t = 1 Leo Storme Constant dimension subspace codes

  17. I MPROVEMENT TO UPPER BOUND FOR t = 1 Eventually PG ( 2 k − 2 , q ) = � c , P 1 , . . . , P k − 1 � ⊂ S . But |S| > | PG ( 2 k − 2 , q ) | . |S \ PG ( 2 k − 2 , q ) | ≈ q 2 k − 3 . Leo Storme Constant dimension subspace codes

  18. I MPROVEMENT TO UPPER BOUND FOR t = 1 If point of S \ PG ( 2 k − 2 , q ) in q k − 1 q − 1 codewords, then PG ( 2 k − 1 , q ) ⊂ S . (FALSE) So all points of S \ PG ( 2 k − 2 , q ) in less than q k − 1 q − 1 codewords. Leo Storme Constant dimension subspace codes

  19. I MPROVEMENT TO UPPER BOUND FOR t = 1 Leo Storme Constant dimension subspace codes

  20. I MPROVEMENT TO UPPER BOUND FOR t = 1 All points of S \ PG ( 2 k − 2 , q ) in less than q k − 1 q − 1 codewords. |S \ PG ( 2 k − 2 , q ) | ≈ q 2 k − 3 . So number of points of S in q k − 1 q − 1 codewords, is approximately q 2 k − 4 . (TOO SMALL) Leo Storme Constant dimension subspace codes

  21. I MPROVEMENT TO UPPER BOUND FOR t = 1 T HEOREM (B ARTOLI , R IET , S TORME , V ANDENDRIESSCHE ) Every 1-intersecting constant dimension code C of codewords of dimension k of size � q k − q � 2 � q k − q � | C | = + + 1 − δ, q − 1 q − 1 with δ = q k − 2 , is sunflower. Leo Storme Constant dimension subspace codes

  22. Thank you very much for your attention! Leo Storme Constant dimension subspace codes

Recommend


More recommend