top of page

Algorithmic Rearrangements of the Alternating Harmonic Series

Aditya Srivastava
31/10/2026

Riemann’s Rearrangement Theorem shows that any conditionally convergent series can be rearranged to sum to any target real number. However, standard constructive proofs do not detail the structural characteristics of the block patterns required for prescribed numerical tolerances. In this paper, I develop a Python implementation using continued fractions to select static block patterns—alternating between p positive terms and n negative terms—that approximate target sums for the alternating harmonic series. I evaluate these rearrangements by tracking selected block sizes and block balance metrics C=121p−1n, comparing algorithmically selected convergents with minimal admissible pairs across different target regions. The results demonstrate that while exact clean representations exist for clean analytical logarithmic targets (e.g., a perfect 1:2 ratio for 12ln2), nearby rounded decimal target values require vastly larger block sizes (such as p=9466,n=18931 at S=0.3466). This happens because these targets are just outside the tolerance window around 12ln2, and stopping at full convergents makes the blocks even larger (the smallest valid pair is 6867:13733). Furthermore, negative targets require heavily imbalanced blocks (pn), driving C higher. Overall, this work establishes a quantitative framework connecting Riemann rearrangements, rational approximation, and block size structural attributes.

bottom of page