combinatorics / Discrete and computational geometry; k-sets
A power saving over Dey's bound for planar halving lines
Theorem 1.1: there are absolute ε>0, C, n0 with h(P)≤Cn4/3−ε for every even n≥n0 and every n-point planar set with no three collinear. Theorem 1.2 bounds the switches at every rank uniformly by CN4/3−ε, and Corollary 1.3 gives O(n(k+1)1/3−ε0)k-sets for 1≤k≤n/2 (also for levels in line arrangements). The constants and exponent are ineffective; no numerical value of ε is given. It does not determine the order of h(n), which could still lie anywhere between neΩ(logn) and n4/3−ε.
Theorem 1.1: there are absolute ε>0, C, n0 with h(P)≤Cn4/3−ε for every even n≥n0 and every n-point planar set with no three collinear. Theorem 1.2 bounds the switches at every rank uniformly by CN4/3−ε, and Corollary 1.3 gives O(n(k+1)1/3−ε0)k-sets for 1≤k≤n/2 (also for levels in line arrangements). The constants and exponent are in…
Theorem 1.1: there are absolute ε>0, C, n0 with h(P)≤Cn4/3−ε for every even n≥n0 and every n-point planar set with no three collinear. Theorem 1.2 bounds the switches at every rank uniformly by CN4/3−ε, and Corollary 1.3 gives O(n(k+1)1/3−ε0)k-sets for 1≤k≤n/2 (also for levels in line arrangements). The constants and exponent are ineffective; no numerical value of ε is given. It does not determine the order of h(n), which could still lie anywhere between neΩ(logn) and n4/3−ε.
Theorem 1.1: there are absolute ε>0, C, n0 with h(P)≤Cn4/3−ε for every even n≥n0 and every n-point planar set with no three collinear. Theorem 1.2 bounds the switches at every rank uniformly by CN4/3−ε, and Corollary 1.3 gives O(n(k+1)1/3−ε0)k-sets for 1≤k≤n/2 (also for levels in line arrangements). The constants and exponent are ineffective; no numerical value of ε is given. It does not determine the order of h(n), which could still lie anywhere between neΩ(logn) and n4/3−ε.