A Simplified Method for Optimising Geometrically Shaped Constellations of Higher Dimensionality

TL;DR

This paper presents a simplified approach for optimizing high-dimensional geometric constellations, achieving significant performance gains over existing methods by designing up to 12-dimensional constellations with 4096 points.

Contribution

A new simplified loss function calculation method enables efficient optimization of high-dimensional constellations, leading to improved performance.

Findings

Designed up to 12D constellations with 4096 points

Achieved gains up to 0.31 dB over state-of-the-art

Demonstrated effectiveness of the simplified optimization method

Abstract

We introduce a simplified method for calculating the loss function for use in geometric shaping, allowing for the optimisation of high dimensional constellations. We design constellations up to 12D with 4096 points, with gains up to 0.31 dB compared to the state-of-the-art.