karthik/Calculations
1
0
Fork 0
Code Issues Pull Requests repo.project_board Releases Wiki Activity

FULLY WORKING VERSION ACHIEVED - fixed all remaining bugs, adjusted parameters for heavy ball method. Both now return density modulated states similar to Santo Roccuzzo's result.

Browse Source
This commit is contained in:
Karthik 2025-04-29 22:11:29 +02:00
parent 7e5aa3df23
commit 055ca7ddc9
2 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
Dipolar-Gas-Simulator/+Scripts/run_locally.m
View File

@ -20,7 +20,7 @@ OptionsStruct.UseApproximationForLHY = true;
OptionsStruct.IncludeDDICutOff = true;
OptionsStruct.CutoffType = 'Cylindrical';
OptionsStruct.SimulationMode = 'EnergyMinimization'; % 'ImaginaryTimeEvolution' | 'RealTimeEvolution' | 'EnergyMinimization'
OptionsStruct.GradientDescentMethod = 'NonLinearCGD'; % 'HeavyBall' | 'NonLinearCGD'
OptionsStruct.GradientDescentMethod = 'HeavyBall'; % 'HeavyBall' | 'NonLinearCGD'
OptionsStruct.MaxIterationsForGD = 2E5;
OptionsStruct.TimeStepSize = 1E-4; % in s
OptionsStruct.MinimumTimeStepSize = 2E-10; % in s

4
Dipolar-Gas-Simulator/+Simulator/@DipolarGas/runGradientDescent.m
View File

@ -8,9 +8,9 @@ function [psi] = runGradientDescent(this,psi,Params,Transf,VDk,V,Observ)
switch this.GradientDescentMethod
case 'HeavyBall'
% Convergence Criteria:
alpha = 1E-6;
alpha = 1E-3;
beta = 0.9;
epsilon = 1E-8;
epsilon = 1E-6;
Observ.residual = 1;
Observ.res = 1;