Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated


Claim. P. Erdős, A. Hajnal and Zs. Tuza, Local constraints ensuring small representing sets, J. Combin. Theory Ser. A 58 (1991), no. 1, 78--84, write (p,1)→rt(p,1)\to_r t when every rr-uniform set system whose subsystems on at most pp elements can each be covered by one element can itself be covered by tt elements, and t0(p,r)t_0(p,r) for the least such tt. The best tt of Problem 616 is t0(3r−3,r)t_0(3r-3,r). Their Theorem 3 (p. 80) proves the upper bound

(3r−3,1)→r⌈r/5⌉(r≥3),(3r-3,1)\to_r\lceil r/5\rceil \qquad (r\ge3),

as a particular case of their Theorem 6 (Section 4, pp. 83--84; the check that p(r,⌈r/5⌉)≤3r−3p(r,\lceil r/5\rceil)\le3r-3 is on p. 84). The prose after Theorem 3 states the lower bound t0(3r−3,r)≥⌊316r+78⌋t_0(3r-3,r)\ge\lfloor\frac3{16}r+\frac78\rfloor, from the set systems H(r,k,q)\mathscr H(r,k,q) of Section 3, and p. 84 gives the construction: with x=⌊316r−18⌋x=\lfloor\frac3{16}r-\frac18\rfloor, q=2x+1q=2x+1 and k=3x+1k=3x+1, every subsystem of H(r,k,q)\mathscr H(r,k,q) on at most 3r−33r-3 elements is covered by one element, while the whole system has covering number k−q+1=x+1=⌊316r+78⌋k-q+1=x+1=\lfloor\frac3{16}r+\frac78\rfloor. So

⌊316r+78⌋≤t≤⌈r5⌉(r≥3).\Bigl\lfloor\tfrac3{16}r+\tfrac78\Bigr\rfloor\le t\le\Bigl\lceil\tfrac r5\Bigr\rceil \qquad (r\ge3).

Covers. The value t=⌈r/5⌉t=\lceil r/5\rceil for exactly the rr at which the two bounds meet: r=3r=3--1010, 1212--1515, 1717--2020, 2222--2525, 2828--3030, 3333--3535, 3838--4040, 4444, 4545, 4949, 5050, 5454, 5555, 6060, 6565 and 7070, thirty-eight values in all, with t=1t=1 for r≤5r\le5, t=2t=2 for 6≤r≤106\le r\le10, and so on up to t=14t=14 at r=70r=70. For every other rr, and so for every r>70r>70, the bounds differ and the best tt is not determined.

Depends on. Nothing in this wiki; the proofs are the paper's own.

Acceptance. Refereed: Journal of Combinatorial Theory, Series A 58 (1991), no. 1, 78--84, received 2 July 1989; the publisher's record dates the issue September 1991, and the page's date is the month's first day. The site labels the problem OPEN and credits the paper only with the two bounds, so reviewed is not listed. No independent proof review and no formalization are recorded.