Home > solver > TEBD_SaveHistory.m

TEBD_SaveHistory

PURPOSE ^

Save history: Save the progress at the current iteration

SYNOPSIS ^

function runhist = TEBD_SaveHistory(obj,runhist,iter,residuals,secs,lambda_ratio,theta)

DESCRIPTION ^

Save history: Save the progress at the current iteration
%
% runhist = TEBD_SaveHistory(obj,runhist,iter,residuals,secs,lambda_ratio,theta)
%
% input:
%        - obj                : contains the values of the objective function(s)
%        - runhist            : (struct) contains the accumulated running history
%        - iter               : current  iteration
%        - resduals           : (struct) contains the specific residuals of the problem at the current iteration 
%        - secs               : time at the current iteration
%        - lambda_ratio       : ratio between a the adaptive and the non-adaptive stepsizes (lambda)
%        - theta              : value of the dynamic scaling parameter at the current iteration
%        
% output:
%        - runhist                   : (struct) contains the accumulated running history, including the info from the current iteration (iter)
%        - runhist.obj               : accumulated running history of the objective function(s)
%        - runhist.theta             : accumulated running history of the dynamic scaling parameter
%        - runhist.gap               : accumulated running history of the duality gap (if it applies for the given problem)
%        - runhist.resnorm           : accumulated running history of the problem specific norms of the residuals 
%        - runhist.lambda_ratio      : accumulated running history of the ratio between a the adaptive and the non-adaptive stepsizes (lambda)
%        - runhist.residual          : accumulated running history of the specific residuals of the problem at the current iteration 
%        - runhist.cputime           : accumulated running history of the time at every iteration
%        - runhist.finalbestresnorm  : the current minimum norm of the residuals
%        - runhist.objbest           : the objective function(s) at the iteration with minimum norm of the residuals
%        - runhist.bestiteration     : the iteration with minimum norm of the residuals
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% 2EBD-HPE:
% Modified by C.Ortiz
% Last Modified: 6/15/2012

CROSS-REFERENCE INFORMATION ^

This function calls: This function is called by:

SOURCE CODE ^

0001 %Save history: Save the progress at the current iteration
0002 %%
0003 %% runhist = TEBD_SaveHistory(obj,runhist,iter,residuals,secs,lambda_ratio,theta)
0004 %%
0005 %% input:
0006 %%        - obj                : contains the values of the objective function(s)
0007 %%        - runhist            : (struct) contains the accumulated running history
0008 %%        - iter               : current  iteration
0009 %%        - resduals           : (struct) contains the specific residuals of the problem at the current iteration
0010 %%        - secs               : time at the current iteration
0011 %%        - lambda_ratio       : ratio between a the adaptive and the non-adaptive stepsizes (lambda)
0012 %%        - theta              : value of the dynamic scaling parameter at the current iteration
0013 %%
0014 %% output:
0015 %%        - runhist                   : (struct) contains the accumulated running history, including the info from the current iteration (iter)
0016 %%        - runhist.obj               : accumulated running history of the objective function(s)
0017 %%        - runhist.theta             : accumulated running history of the dynamic scaling parameter
0018 %%        - runhist.gap               : accumulated running history of the duality gap (if it applies for the given problem)
0019 %%        - runhist.resnorm           : accumulated running history of the problem specific norms of the residuals
0020 %%        - runhist.lambda_ratio      : accumulated running history of the ratio between a the adaptive and the non-adaptive stepsizes (lambda)
0021 %%        - runhist.residual          : accumulated running history of the specific residuals of the problem at the current iteration
0022 %%        - runhist.cputime           : accumulated running history of the time at every iteration
0023 %%        - runhist.finalbestresnorm  : the current minimum norm of the residuals
0024 %%        - runhist.objbest           : the objective function(s) at the iteration with minimum norm of the residuals
0025 %%        - runhist.bestiteration     : the iteration with minimum norm of the residuals
0026 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
0027 %% 2EBD-HPE:
0028 %% Modified by C.Ortiz
0029 %% Last Modified: 6/15/2012
0030 function runhist = TEBD_SaveHistory(obj,runhist,iter,residuals,secs,lambda_ratio,theta)
0031 
0032 runhist.obj{iter}= obj;
0033 
0034 runhist.theta(iter)= theta;
0035 
0036 runhist.gap(iter)= residuals.gap;
0037 
0038 runhist.resnorm(iter) = residuals.resnorm;
0039 
0040 runhist.lambda_ratio(iter) = lambda_ratio;
0041 
0042 runhist.residual(iter) = residuals;
0043 
0044 runhist.cputime(iter) = secs;
0045 
0046 if iter == 1 || runhist.finalbestresnorm >= residuals.resnorm
0047     runhist.finalbestresnorm = residuals.resnorm;
0048     runhist.objbest(1)=runhist.obj(iter);
0049     runhist.bestiteration = iter;
0050 end
0051 
0052

Generated on Tue 06-Aug-2013 17:07:59 by m2html © 2005