1 / 58

A new family of expander Cayley graphs (?)

A new family of expander Cayley graphs (?). Eyal Rozenman, Hebrew University Aner Shalev, Hebrew University Avi Wigderson, IAS You need the TexPoint PowerPoint add-in to view this presentation properly Download it from http://raw.cs.berkeley.edu/texpoint/. Definitions.

avital
Download Presentation

A new family of expander Cayley graphs (?)

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. A new family of expander Cayley graphs (?) Eyal Rozenman, Hebrew University Aner Shalev, Hebrew University Avi Wigderson, IAS You need the TexPoint PowerPoint add-in to view this presentation properly Download it from http://raw.cs.berkeley.edu/texpoint/

  2. Definitions Undirected, regular (multi)graphs. Definition. The 2nd eigenvalue of a d-regular X (X) = max { || (AX /d) v || : ||v||=1 , v  1 } (X)  [0,1] Definition. {Xn}is an expander family if (Xn) <1 Equivalent: RW on X converges in time O(log|X|)

  3. Cayley graphs • Ga finite group • U½ Ga (symmetric) set of generators. • Definition: Cayley graph C(G,U): • Vertices: elements of G • Edges :(g, gu)for all uU. • C(G,U)is regular with degree |U|. • C(G,U) is connected ,U generates G.

  4. 1 2 3 4 1 Our Sequence of groups Gn = Even symmetries of a (rooted) d-regular tree. degree (#children) = d (fixed). Depth = n

  5. 1 2 3 4 1 Our Sequence of groups Gn = Even symmetries of a (rooted) d-regular tree. degree (#children) = d (fixed). Depth = n Depth 1: tree symmetries = alternating groupAd

  6. 1 2 3 4 1 Our Sequence of groups Gn = Even symmetries of a (rooted) d-regular tree. degree (#children) = d (fixed). Depth = n Depth 1: tree symmetries = alternating groupAd Theorem: Under some assumption on Ad: Every Gn has (const. #) expanding generators.

  7. Assumption on alternating gp. Ad Our construction is based on Assumption:9U½Ad (= G1) such that • |U| ·d1/30 • (Ad,U) ·1/1000 This is an Open problem (will discuss this more later)

  8. 1 2 3 4 1 Main theorem Thm [RSW]: 9Un½Gn such that • |Un| ·d1/7 (constant -independent of n) • C(Gn,Un) is a good expander ((GnUn) · 1/1000)

  9. 1 2 3 4 1 Main theorem Thm [RSW]: 9Un½Gn such that • |Un| ·d1/7 (constant -independent of n) • C(Gn,Un) is a good expander ((GnUn) · 1/1000) Proof is by induction on n (base of induction is the assumption on Ad)

  10. Ad Gn Gn Gn Gn Inductive definition of groups G1 = Ad, the alternating group. • Gn+1 = GnoAd (wreath product) • Gn+1= {(,x1,x2,…,xd) | 2Ad,xi2G }

  11. Ad Gn Gn Gn Gn Inductive definition of groups G1 = Ad, the alternating group. • Gn+1 = GnoAd (wreath product) • Gn+1= {(,x1,x2,…,xd) | 2Ad,xi2G } Multiplication rule: (,x1,x2,…,xd) ¢ (,y1,y2,…,yd) = (, x(1),y1, x (2)y2,…, x (d)yd)

  12. Ad Gn Gn Gn Gn Inductive definition of groups G1 = Ad, the alternating group. • Gn+1 = GnoAd (wreath product) • Gn+1= {(,x1,x2,…,xd) | 2Ad,xi2G } Multiplication rule: (,x1,x2,…,xd) ¢ (,y1,y2,…,yd) = (, x(1),y1, x (2)y2,…, x (d)yd) Example: conjugation g=(,1,…,1) , x = (1,x1,…,xd) g-1 x g=(1,x(1),…,x(d))

  13. Generating sets U1½Ad – “small” generating set Un½Gn – “small” generating set Goal: Un+1½Gn+1 – “small” generating set

  14. Generating sets U1½Ad – “small” generating set Un½Gn – “small” generating set Goal: Un+1½Gn+1 – “small” generating set Embed U1½Gn : {(u,1,1,L,1) |u2U1} Generates all {(s,1,1,L,1) |s2Ad} U1½Ad 12Gn 12Gn 12Gn 12Gn

  15. Generating sets Pick some x=(1,x1,x2,L,xd) 2Gn+1 1 x12Gn x22Gn xd-12Gn xd2Gn

  16. Generating sets Pick some x=(1,x1,x2,L,xd) 2Gn+1 Conjugating x by {(s,1,1,L,1) permutes xby s We get all (even)permutations of x. 1 x12Gn x22Gn xd-12Gn xd2Gn

  17. Generating sets Pick some x=(1,x1,x2,L,xd) 2Gn+1 Conjugating x by {(s,1,1,L,1) permutes xby s We get all (even)permutations of x. If the permutations of xgenerate (Gn)d Then U1[ {x} generates Gn+1 1 x12Gn x22Gn xd-12Gn xd2Gn

  18. Generating sets Necessary condition: {x1,x2,L,xd} generates Gn 1 x12Gn x22Gn xd-12Gn xd2Gn

  19. Generating sets Necessary condition: {x1,x2,L,xd} generates Gn Un generatesGn. Un={u1,u2,,up}. Suppose p dividesd Put x = (u1,u1,,u1, u2,u2,,u2,, up,up,,up) 1 x12Gn x22Gn xd-12Gn xd2Gn

  20. Generating sets Necessary condition: {x1,x2,L,xd} generates Gn Un generatesGn. Un={u1,u2,,up}. Suppose p dividesd Put x = (u1,u1,,u1, u2,u2,,u2,, up,up,,up) DefineUn(d) = the orbit of x under Ad. 1 x12Gn x22Gn xd-12Gn xd2Gn

  21. Generating sets Necessary condition: {x1,x2,L,xd} generates Gn Un generatesGn. Un={u1,u2,,up}. Suppose p dividesd Put x = (u1,u1,,u1, u2,u2,,u2,, up,up,,up) DefineUn(d) = the orbit of x under Ad. Does Un(d) generate (Gn)d ? Is it expanding ? 1 x12Gn x22Gn xd-12Gn xd2Gn

  22. Algebraic Zig-Zag theorem [ALW] If • ( (Gn)d, Un(d) ) · 1/50 • (Ad,U1) · 1/1000

  23. Algebraic Zig-Zag theorem [ALW] If • ( (Gn)d, Un(d) ) · 1/50 • (Ad,U1) · 1/1000 Then 9W½Gn+1(= GnoAd) • (Gn+1, W) · 1/50 + 1/1000

  24. Algebraic Zig-Zag theorem [ALW] If • ( (Gn)d, Un(d) ) · 1/50 • (Ad,U1) · 1/1000 Then 9W½Gn+1(= GnoAd) • (Gn+1, W) · 1/50 + 1/1000 • |W| = |U1|2 (Important! |W| is a func. of |U1|=const.)

  25. Proof of main theorem Induction assumption: 9Un½Gn , (Gn,Un) · 1/1000

  26. Proof of main theorem Induction assumption: 9Un½Gn , (Gn,Un) · 1/1000 Main Lemma:( (Gn)d, Un(d) ) · 1/50 (for good G,U)

  27. Proof of main theorem Induction assumption: 9Un½Gn , (Gn,Un) · 1/1000 Main Lemma:( (Gn)d, Un(d) ) · 1/50 (for good G,U) Zigzag thm [RVW,ALW]: 9W½Gn+1(= GnoAd) • (Gn+1, W) · 1/50 + 1/1000 • |W| = |U1|2

  28. Proof of main theorem Induction assumption: 9Un½Gn , (Gn,Un) · 1/1000 Main Lemma:( (Gn)d, Un(d) ) · 1/50 (for good G,U) Zigzag thm [RVW,ALW]: 9W½Gn+1(= GnoAd) • (Gn+1, W) · 1/50 + 1/1000 • |W| = |U1|2 DefineUn+1 = W*2 := all words of length 2 in W • |Un+1| = |U1|4·d1/7 • (Gn+1, Un+1) · (1/50 + 1/1000)2· 1/50

  29. Main Lemma Main Lemma: Given • U½G • (G, U)· 1/1000 then - ( (G)d, U(d)) · 1/50 (for good G,U)

  30. Main Lemma Main Lemma: Given • U½G • (G, U)· 1/1000 then - ( (G)d, U(d)) · 1/50 (for good G,U) Example: ( (G)d, Ud ) · 1/1000 But: Ud is NOT one Ad orbit

  31. Main Lemma Main Lemma: Given • U½G small enough • (G, U)· 1/1000 then - ( (G)d, U(d)) · 1/50 (for good G,U) Example: ( (G)d, Ud ) · 1/1000 But: Ud is NOT one Ad orbit The idea – when |U|<< d • A random element of Ud is “more or less” in U(d). • So U(d) ”approximates” Ud well.

  32. Main Lemma-reduction to G £ G When/Why is C( Gd, U(d)) connected ?

  33. Main Lemma-reduction to G £ G When/Why is C( Gd, U(d)) connected ? (u1 , u2 , u3 , L,ud) 2U(d) (u2-1, u1 –1 , u3-1 , L,ud-1) 2U(d)

  34. Main Lemma-reduction to G £ G When/Why is C( Gd, U(d)) connected ? (u1 , u2 , u3 , L,ud) 2U(d) (u2-1, u1 –1 , u3-1 , L,ud-1) 2U(d) multiply to get (u1, u2-1 ,u2 u1–1 , 1 , L , 1)

  35. Main Lemma-reduction to G £ G When/Why is C( Gd, U(d)) connected ? (u1 , u2 , u3 , L,ud) 2U(d) (u2-1, u1 –1 , u3-1 , L,ud-1) 2U(d) multiply to get (u1, u2-1 ,u2 u1–1 , 1 , L , 1) Setting u2 =1 (assume 1 2 U) we generate all elements { (g , g-1 , 1 , 1, … , 1) | g 2 U} U1

  36. Main Lemma-reduction to G £ G When/Why is C( Gd, U(d)) connected ? (u1 , u2 , u3 , L,ud) 2U(d) (u2-1, u1 –1 , u3-1 , L,ud-1) 2U(d) multiply to get (u1, u2-1 ,u2 u1–1 , 1 , L , 1) Setting u2 =1 (assume 1 2 U) we generate all elements { (g , g-1 , 1 , 1, … , 1) | g 2 U} • If first two coordinates generate G£G we are done • Expansion also follows U1

  37. Expansion on G £ G X½G Z½G£G Z = {(x,x-1) | x2X} (completely correlated!)

  38. Expansion on G £ G X½G Z½G£G Z = {(x,x-1) | x2X} (completely correlated!) Suppose (G,X) · 1- Is (G£G, Z) · 1-f() for some f?

  39. Expansion on G £ G X½G Z½G£G Z = {(x,x-1) | x2X} (completely correlated!) Suppose (G,X) · 1- Is (G£G, Z) · 1-f() for some f? G abelian – NO. • C (G£G, Z) is disconnected. • Connected component of (1,1) is {(g,g-1) | g2G}

  40. Expansion on G £ G X½G Z½G£G Z = {(x,x-1) | x2X} (completely correlated!) Suppose (G,X) · 1- Is (G£G, Z) · 1-f() for some f? G abelian – NO. • C (G£G, Z) is disconnected. • Connected component of (1,1) is {(g,g-1) | g2G} If Y = {(x,x) | x2X} then C(G£G, Y) is disconnected

  41. Expansion on G £ G – decorrelating the gens. BUT: If for every x2G, x = [ax,bx] = ax-1 bx-1 ax bx

  42. Expansion on G £ G – decorrelating the gens. BUT: If for every x2G, x = [ax,bx] = ax-1 bx-1 ax bx Xc = [ {ax,bx,(ax-1¢bx-1) | x2X} (+ inverses) Z½G£G Z = {(x,x-1) | x2Xc}

  43. Expansion on G £ G – decorrelating the gens. BUT: If for every x2G, x = [ax,bx] = ax-1 bx-1 ax bx Xc = [ {ax,bx,(ax-1¢bx-1) | x2X} (+ inverses) Z½G£G Z = {(x,x-1) | x2Xc} (ax-1bx-1 , bxax) 2Z (ax , ax-1) 2Z (bx , bx-1) 2Z

  44. Expansion on G £ G – decorrelating the gens. BUT: If for every x2G, x = [ax,bx] = ax-1 bx-1 ax bx Xc = [ {ax,bx,(ax-1¢bx-1) | x2X} (+ inverses) Z½G£G Z = {(x,x-1) | x2Xc} (ax-1bx-1 , bxax) 2Z (ax , ax-1) 2Z (bx , bx-1) 2Z + (x , 1) is generated by Z

  45. Expansion on G £ G – decorrelating the gens. BUT: If for every x2G, x = [ax,bx] = ax-1 bx-1 ax bx Xc = [ {ax,bx,(ax-1¢bx-1) | x2X} (+ inverses) Z½G£G Z = {(x,x-1) | x2Xc} (ax-1bx-1 , bxax) 2Z (ax , ax-1) 2Z (bx , bx-1) 2Z + (x , 1) is generated by Z If (G,X) · 1- then (G£G, Z) · 1-( /500|X|2)

  46. Commutator representation in Gn Def a group G has the commutator property (CP) if every element in G is a commutator.

  47. Commutator representation in Gn Def a group G has the commutator property (CP) if every element in G is a commutator. Theorem [ORR, 50s] Ad has (CP)

  48. Commutator representation in Gn Def a group G has the commutator property (CP) if every element in G is a commutator. Theorem [ORR, 50s] Ad has (CP) Theorem [N 03’] If G has (CP) then G oAd has (CP)

  49. Commutator representation in Gn Def a group G has the commutator property (CP) if every element in G is a commutator. Theorem [ORR, 50s] Ad has (CP) Theorem [N 03’] If G has (CP) then G oAd has (CP) Proof: Reduce to the system of eqs x1 yh(1) xs(1)-1 yt(1) = a1 x2 yh(2) xs(2)-1 yt(2) =a2 … xd yh(d) xs(d)-1 yt(d) =ad 2d variables xi2G, yi2G, d constants ai2G. 3 arbitrary permutations h,s,t2Sd

  50. Expansion on alternating gp. Ad Our construction was based on Assumption:9U½Ad (= G1) such that • |U| ·d1/30 • (Ad,U) ·1/1000

More Related