Some of the properties of the sequence of powers of prime numbers. (Q1414927): Difference between revisions

For every positive real number \(x\) let \(\pi^*(x)\) denote the number of prime powers \(p^{\alpha}\leq x\), and for all integers \(n\geq 1\) let \(q_n\) denote the \(n\)th prime power. In this paper, the author shows that the behaviour of the functions \(\pi^*(x)\) and \(q_n\) resemble the behaviour of the functions \(\pi(x)\) and \(p_n\) giving the number of primes \(p\leq x\) and the \(n\)th prime, respectively. For example, the author proves (Theorem 1) that for every natural number \(n\) and a suitable choice of signs \(\pm \), one has \(q_{n+1}=q_n\pm q_{n-1} \pm \dots\pm q_1+\varepsilon_n\), where \(\varepsilon_n\in \{0,1\}\), the analogous result with \(q_n\) replaced by \(p_n\) being a result of Sierpiński. The author also proves that \(\pi^*(2x)\leq \pi^*(x)\) holds for all integer values of \(x\geq 3\) and that the quantity \(q_{n+1}-2q_n+q_{n-1}\) changes signs infinitely often. The proofs are elementary modulo the use of some prior improvements, due to the author, of results of Rosser and Schoenfeld regarding estimates with explicit error terms for the function \(\pi(x)\). The paper concludes with a couple of open questions.