The Chung-Graham Expansion
Sungkon Chang
TL;DR
This work generalizes the Chung-Graham expansion by showing that every positive integer admits a unique expansion using Fibonacci terms with equally spaced indices $F_{2+d(k-1)}$, for fixed even $d$. It defines the Chung-Graham rule with bounds $A=K_d-1$ and $B=F_{2+d}-1$, proves existence and uniqueness of expansions via a constructive induction and a lexicographic-block approach, and develops a detailed combinatorial framework of coefficient sequences and blocks. Central to the result is a maximal coefficient sequence $\beta^{(n)}$ and a block-decomposition theory that yields a well-defined, order-based construction of all feasible coefficient sequences; this leads to a bijection between coefficient sequences and positive integers through the evaluation map. The findings extend Zeckendorf-type representations to broader Fibonacci-indexed subsequences and provide a robust, constructive method for obtaining unique representations with equally-spaced Fibonacci terms, with potential applications to financial series, combinatorial encoding, and number representations.
Abstract
Chung and Graham introduced a method to uniquely represent each positive integer using even-indexed Fibonacci terms. We generalize this result to represent each positive integer using other Fibonacci terms with equally-spaced indices.
