Collective Coordinate Control of Density Distributions

TL;DR

This paper develops a numerical optimization method to generate many-particle configurations with prescribed collective density variables and structure factors, extending previous studies to three dimensions and near-zero wave vectors, with applications in physics and cosmology.

Contribution

It introduces an advanced optimization technique to control density distributions in three-dimensional systems, enabling tailored structure factors and exploring degenerate disordered ground states.

Findings

Successfully extended to three dimensions

Controlled structure factor behavior near zero wave vector

Generated configurations with specific power-law structure factors

Abstract

Real collective density variables $C(\boldsymbol{k})$ [c.f. Eq.\ref{Equation3})] in many-particle systems arise from non-linear transformations of particle positions, and determine the structure factor $S(\boldsymbol{k})$, where $\bf k$ denotes the wave vector. Our objective is to prescribe $C({\boldsymbol k})$ and then to find many-particle configurations that correspond to such a target $C({\bf k})$ using a numerical optimization technique. Numerical results reported here extend earlier one- and two-dimensional studies to include three dimensions. In addition, they demonstrate the capacity to control $S(\boldsymbol{k})$ in the neighborhood of $|\boldsymbol{k}| =$ 0. The optimization method employed generates multi-particle configurations for which $S(\boldsymbol{k}) \propto |\boldsymbol{k}|^{\alpha}$, $|\boldsymbol{k}| \leq K$, and $\alpha =$ 1, 2, 4, 6, 8, and 10. The case $\alpha =$ 1 is relevant for the Harrison-Zeldovich model of the early universe, for superfluid $^{4}{He}$, and for jammed amorphous sphere packings. The analysis also provides specific examples of interaction potentials whose classical ground state are configurationally degenerate and disordered.