TLDR: Researchers have developed a new framework combining the PSLQ algorithm with number-theoretic filters to efficiently discover Machin-like arctangent relations for calculating Pi. This approach yielded new 5 and 6-term relations with record-low Lehmer measures, significantly improving the efficiency of Pi computation. The framework also allows for generating longer, more efficient formulae.
For centuries, mathematicians have been fascinated by Pi (Ï€), the fundamental constant representing the ratio of a circle’s circumference to its diameter. Calculating Pi to an ever-increasing number of decimal places has been a long-standing challenge, often relying on elegant mathematical formulas known as Machin-like arctangent relations.
These formulas, first discovered by John Machin in 1706, provide a highly efficient way to compute Pi. The efficiency of these relations is quantified by something called the Lehmer measure. Simply put, a lower Lehmer measure means a more efficient formula for calculating Pi.
A recent research paper by Nick Craig-Wood introduces a groundbreaking new method for discovering these highly efficient Machin-like relations. The core of this method lies in combining a powerful mathematical tool called the PSLQ integer-relation algorithm with clever number-theoretic filters. The PSLQ algorithm is designed to find hidden integer relationships between real numbers, which is crucial for uncovering these complex Pi formulas.
However, using PSLQ alone can be like searching for a needle in a haystack. To make this search manageable on a large scale, the researchers incorporated filters based on the algebraic structure of Gaussian integers. These filters act like a sieve, narrowing down the possibilities and making the search for low-measure relations much more practical.
This innovative approach has led to exciting new discoveries. The search yielded new 5-term and 6-term Machin-like relations that boast record-low Lehmer measures. Specifically, a new 5-term relation achieved a Lehmer measure of 1.4572, and a new 6-term relation achieved an even lower measure of 1.3291. These figures represent the best-known efficiencies for their respective lengths to date, surpassing previous records held for over a century in some cases.
Beyond just finding new formulas, the research also demonstrates how these newly discovered relations can serve as a foundation for generating even longer and potentially more efficient formulas through algorithmic extensions. This combined strategy of a constrained PSLQ search and algorithmic extension offers a robust and promising path for future explorations into the fascinating world of Pi computation.
Also Read:
- Splitting Minds: How Two AI Agents Outperform One in Mathematical Problem Solving
- Enhancing Trust and Performance in Problem Solving with Implicit Hitting Sets
This work not only pushes the boundaries of how we calculate Pi but also showcases a general template for using advanced algorithms like PSLQ to solve complex mathematical problems. For more technical details, you can refer to the full research paper available at arXiv:2508.08307.
