# A practical Seedless Infrared-Safe Cone jet algorithm

**Authors:** Gavin P. Salam, Gregory Soyez

arXiv: 0704.0292 · 2009-11-13

## TL;DR

This paper introduces a new seedless infrared-safe cone jet algorithm that is both exact and computationally efficient, improving the reliability of jet identification in collider experiments.

## Contribution

The paper presents an efficient, exact seedless cone jet algorithm that guarantees IR safety and is practical for hadron collider data analysis.

## Key findings

- The new algorithm is IR safe and identifies all stable cones.
- It operates in N^2 ln(N) time, comparable to existing methods.
- It results in modest changes to inclusive jet spectra and significant changes to multi-jet observables.

## Abstract

Current cone jet algorithms, widely used at hadron colliders, take event particles as seeds in an iterative search for stable cones. A longstanding infrared (IR) unsafety issue in such algorithms is often assumed to be solvable by adding extra `midpoint' seeds, but actually is just postponed to one order higher in the coupling. A proper solution is to switch to an exact seedless cone algorithm, one that provably identifies all stable cones. The only existing approach takes N 2^N time to find jets among N particles, making it unusable at hadron level. This can be reduced to N^2 ln(N) time, leading to code (SISCone) whose speed is similar to that of public midpoint implementations. Monte Carlo tests provide a strong cross-check of an analytical proof of the IR safety of the new algorithm, and the absence of any 'R_{sep}' issue implies a good practical correspondence between parton and hadron levels. Relative to a midpoint cone, the use of an IR safe seedless algorithm leads to modest changes for inclusive jet spectra, mostly through reduced sensitivity to the underlying event, and significant changes for some multi-jet observables.

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/0704.0292/full.md

## Figures

19 figures with captions in the complete paper: https://tomesphere.com/paper/0704.0292/full.md

## References

38 references — full list in the complete paper: https://tomesphere.com/paper/0704.0292/full.md

---
Source: https://tomesphere.com/paper/0704.0292