On Kaprekar's Junction Numbers

Abstract: A base-b junction number u has the property that there are at least two ways to write it as u = v + s(v), where s(v) is the sum of the digits in the expansion of the number v in base b. For the base-10 case, Kaprekar in the 1950's and 1960's studied the problem of finding K(n), the smallest u such that the equation u=v+s(v) has exactly n solutions. He gave the values K(2)=101, K(3)=10^13+1, and conjectured that K(4)=10^24+102. In 1966 Narasinga Rao gave the upper bound 10^1111111111124+102 for K(5), as well as upper bounds for K(6), K(7), K(8), and K(16). In the present work, we derive a set of recurrences which determine K(n) for any base b and in particular imply that these conjectured values of K(n) are correct. The key to our approach is an apparently new recurrence for F(u), the number of solutions to u=v+s(v). We illustrate our method by computing the values of K(n) for n <= 16 and bases b <= 10, and show that for each base K(n) grows as a tower of height proportional to log_2(n). Rather surprisingly, the values of K(n) for the base-5 problem are determined by the classical Thue-Morse sequence, which leads us to define generalized Thue-Morse sequences for other bases.