On perfect and near-perfect numbers (Q455810)

Let \(\sigma(n)\) denote the sum of the positive divisors of \(n\). The integer \(n\) is perfect if \(\sigma(n)= 2n\), and there are many unanswered questions concerning such integers. The current authors consider near-perfect integers \(n\) given by \(\sigma(n)= 2n+ d\) where \(d\) is a proper divisor of \(n\) called a redundant divisor. These are a special type of pseudo-perfect number as defined by Sierpiński. They verify that the following integers \(n\) are near-perfect: (i) \(n= 2^{t-1}(2^t- 2^k- 1)\) where \(t\geq k+ 1\) and \(2^t- 2^k- 1\) is prime with \(2^k\) the redundant divisor; (ii) \(n= 2^{p-1}(2^p- 1)^2\) where \(p\) and \(2^p- 1\) are primes with \(2^p- 1\) the redundant divisor, so \(n\) is derived from the perfect number \(2^{p-1}(2^p- 1)\). The authors remark that whether there are infinitely many near-perfect numbers is yet to be established. Their main result asserts that the number of near-perfect numbers \(n\leq x\) is at most \(x^{{5\over 6}+o(1)}\) as \(x\to\infty\). Several lemmas are applied in the proof, including one deriving an upper bound uniform in \(a\) for the number of sporadic solutions \(n\) of the congruence \(\sigma(n)\equiv a\pmod n\), that is those solutions \(n\) not of the form \(n= pm\) with \(p\) prime, \(p\nmid m\), \(m\mid\sigma(m)\) and \(\sigma(m)= a\). When \(a\) is fixed, \textit{C. Pomerance} [Acta Arith. 26, 265--272 (1975; Zbl 0266.10005)] derived an upper bound for this number of sporadic solutions, and the present authors establish a stronger result uniform in \(a\). In the final section they derive results concerning \(k\)-near-perfect integers \(n\) where the single redundant divisor \(d\) above is replaced by a sum of \(k\) proper divisors of \(n\), so \(\sigma(n)= 2n+ d_1+\cdots+ d_k\) where \(d_1<\cdots< d_k\) are proper divisors of \(n\). Their main result here is that if \(k\) is large enough there are infinitely many \(k\)-near-perfect numbers \(n\).