On the number of partitions into primes (Q2426702)

As is implicit in the author's introductory discussion (\S 1), what constitutes a bona fide asymptotic formula depends uperness expectations concerning the main term. Here, the dominant term is reasonably uncomplicated, while at same time possessing the two advantages mentioned in the last line of the author's abstract. In this sense, the expression for the main term compares favorably with the more general results of \textit{K. F. Roth} and \textit{G. Szekeres} [Q. J. Math., Oxf. II. Ser. 5, 241--259 (1954; Zbl 0057.03902)], who obtain an asymptotic formula for the number of partitions of a nonnegative integer into elements of an eventually increasing sequence of integers. The main result of the paper in Theorem 2, which may be stated as follows. For \(n\in\mathbb{Z}^+\) let \(P(n)\) denote the number of partitions of \(n\) into prime numbers; that is, \(P(n)\) is the number of solutions of \(k_1p_1+k_2p_2+\cdots=n\), with \(k_j\in\mathbb{Z}^+\) and \(\{p_j\}\) the sequence of primes in increasing order. Put \(\Phi(z)=\sum^\infty_{n=0}P(n)z^n=\prod_p(1-z^p)^{-1}\); \(\Phi(z)\) is holomorphic and \(\neq 0\) in \(D=\{z:| z|<1\}\). For \(z\) in \(D\), \(\Phi(z)=\exp(\Psi(z))\), where \(\Psi(z)=\sum^\infty_{k=1} \sum_p\frac{z^kp}{k}\). Also, for real \(x\in 0\), the equation \(\rho\psi' (p)=x\) has a unique solution \(\rho=\rho(x)\), with \(0\leq \rho<1\) and \(\rho(x)\to 1^-\) as \(x\to+\infty\). Then, for \(n\to+\infty\), \[ P(n)=\frac {\rho(n)^{-n}\Phi(\rho(n))}{\sqrt{2\pi\psi_2(\rho(n))}}(1+\theta(n^{-1/5})),\tag{*} \] where \(\Psi_2(\rho)(\rho\frac{d}{d\rho})^2\Psi(\rho)\). A related result is Theorem 3. As \(n\to\infty\), \(P(n+1)-P(n)\sim\frac{\pi P(n)}{\sqrt{2n\log n}}\). The author points out that ``the error estimate in Theorem 2 is inadequate for anything useful to be deduced about \(P(n+1)-P(n)\). However the method of proof of Theorem 2 can be adapted to obtain Theorem 3. The proof of Theorem 2 is an impressive application of the circle method in the great tradition established by Hardy and Ramanujan, and continued in the present day by masters of the estimable art of estimation. On this article Robert Vaughan shows once again that he has a place in this elite group.