Home 9 Science 9 New Progress on a Longstanding Wave Problem

New Progress on a Longstanding Wave Problem

by | Jan 30, 2026

Mathematicians leverage network theory to untangle a decades-old Fourier mystery.
Source: Samuel Velasco/Michael Kanyongolo/Quanta Magazine.

 

Mathematicians have made a significant breakthrough on a problem about wave sums that resisted progress for decades. The challenge, known as Chowla’s cosine problem, asks how low a sum of simple cosine waves can go when constructed from a set of integer frequencies. First posed in 1965, it became a touchstone for understanding limitations in harmonic analysis and the Fourier transform, a foundational tool in mathematics and engineering that represents functions as sums of waves, tells Quanta Magazine.

For years, researchers made only incremental advances. Traditional methods rooted in Fourier techniques, tools that decompose signals into constituent frequencies, failed to yield strong bounds on the minimum values these cosine sums could achieve. Then, in late 2025, a team of four mathematicians, that is, Zhihan Jin, Aleksa Milojević, István Tomon, and Shengtong Zhang, posted a new result that injects fresh life into the decades-old pursuit.

Their approach sidesteps classical Fourier analysis by linking the wave problem to concepts in graph theory and networks. The key connection arises through Cayley graphs, structures that encode relationships between numbers as networks of nodes and edges. The behavior of the smallest eigenvalue of a Cayley graph, a quantity reflecting deep structural properties of the network, corresponds directly to the minimum of the cosine sum under study.

Drawing on recent advances in understanding eigenvalues and MaxCut problems in graphs, the researchers showed that certain network constraints force the cosine sum below new bounds that improve on long-standing estimates. This blending of discrete mathematics and harmonic analysis not only advances Chowla’s problem but also demonstrates how network methods can illuminate classical questions about waves and transformations.

The work highlights a broader trend in mathematics where insights from one area—here, graph structure and spectral properties—unlock progress in another, otherwise stagnant field. It also reinforces the deep ties between networks, signal analysis, and the fundamental mathematics underlying waves.