We present the first explicit comparison-based algorithm that sorts the
sumset $X + Y = \{x_i + y_j,\ \forall 0 \le i, j < n\}$, where $X$ and $Y$ are sorted arrays of real numbers, in optimal $O(n^2)$ time and comparisons. While Fredman (1976) proved the theoretical existence of such an algorithm, a concrete construction has remained open for nearly five decades. Our algorithm exploits the structured monotonicity of the sumset matrix to perform amortized constant-comparisons and insertions, eliminating the $\log(n)$ overhead typical of comparison-based sorting. We prove correctness and optimality in the standard comparison model, extend the method to $k$-fold sumsets with $O(n^k)$ performance, and outline potential support for dynamic updates. Experimental benchmarks show significant speedups over classical algorithms such as MergeSort and QuickSort when applied to sumsets. These results resolve a longstanding open problem in sorting theory and contribute novel techniques for exploiting input structure in algorithm design.

Questo articolo esplora i giri e le loro implicazioni.

Scarica PDF: