On almost primes in Piatetski-Shapiro sequences
Runbo Li
TL;DR
The paper proves that for 0.9985 < \\gamma < 1 there exist infinitely many primes p with \\Omega([p^{1/\\gamma}])\\le 5, improving earlier bounds. It employs Chen's switching principle and Iwaniec's linear sieve in a sieve framework, augmented by two new combinatorial lemmas to control error terms. The main technical achievement is a positive lower bound arising from a carefully balanced weighted sum W(A, x^{1/17.41}) and associated integral estimates, establishing the P5 bound for the almost-prime outputs. The work highlights a barrier to reaching P4 due to limitations in handling higher-dimensional exponential-sum error terms and Richert-type sieve weights, clarifying the method's current cap near P5 and its dependence on γ close to 1.
Abstract
The author proves that for $0.9985 < γ< 1$, there exist infinitely many primes $p$ such that $[p^{1/γ}]$ has at most 5 prime factors counted with multiplicity. This gives an improvement upon the previous results of Banks-Guo-Shparlinski and Xue-Li-Zhang.
