mdapack.m.template

clear variables;
clear import;
close all;
% Configures multiple runs of MDAPACK with differing sets of parameters

% make sure that the model binaries are on the search path
addpath(strcat(pwd,filesep,'bin'));

% Typically unused, can be used to explicitly set the seed to create
% exactly reproducible results
s1=RandStream('mt19937ar','seed',1);
s2=RandStream('mt19937ar','seed',2);
s3=RandStream('mt19937ar','seed',3);

% build averaging Gaussian measurement operator
% width = 2.0;
% mDim = 4;
% nDim = 1024;
% avgGaussH = tools.avgGauss(nDim, mDim, width);

width = 2.0;
mDim = 20;
nDim = 40;
avgGaussH = tools.avgGauss(nDim, mDim, width);
%% 1) Choose representation, model and model parameters
parameters.representation = { ...
%      @(model) representations.Ensemble(model, ...
%      'ensSize', 3 ...
%      ) ... 
    @(model) representations.PCE( ...
      model, ...
     'pceOrder', 3, ...
     'sampleSize', 10000) ... % also used to compute quantiles
};

parameters.model = {... % bind all parameters of the model generator call
%     @() models.RandomWalk('evidenceNoiseRng', s2, 'modelNoiseRng', s1)
%     @() models.Lorenz1996(...
%     'measurementSchedule' , 10:2:120, ... % 10 days mixing, 110 days assimilation
%     'tEnd',                 150, ... % 30 days forecast period
%     'measurementDimension', mDim, ...
%     'measurementStdDev',    0.1, ...
%     'deterministicDimension', nDim ...
%     )
%     'measurementOperator', @(x) models.Lorenz1984.customMeasurementOperator(x, avgGaussH) ...
%     @() models.Lorenz1963(...
%     'tStep', 0.05, ...
%     'measurementSchedule' , 4:2:100, ...
%     'measurementError',    0.5, ...
%     'plot2dDim', 1 ...
%     )...
    @() models.Lorenz1984(...
    'tStep', 0.25, ...
    'measurementSchedule' , 10:2:100, ...
    'measurementStdDev',    0.000000000000000000001, ...
    'plot2dDim', 1, ...
    'plotMode', '2d', ...
    'tEnd', 120, ...
    'rndStream', s1 ...
    )...
%     'measurementOperator', @(x) models.Lorenz1984.customMeasurementOperator(x,[1,1,-1]), ...
%     @() models.Advection(...
%     'deterministicDimension',          1024, ...
%     'decorrelationLength', 20.0,...
%     'measurementDimension',        4, ...
%     'measurementFreq',     5, ...
%     'tCount',              300, ...
%     'measurementStdDev',    0.1, ...
%     'rndStream', s1 ...
%     )...
%     @() models.AdvectionNoise(...
%     'deterministicDimension',          1024, ...
%     'decorrelationLength', 20.0,...
%     'decorrelationLengthSmall', 8.0,...
%     'noiseLevel',          0.04,...
%     'measurementDimension',        4, ...
%     'measurementFreq',     5, ...
%     'tCount',              300, ...
%     'measurementStdDev',    0.1, ...
%     'rndStream', s1 ...
%     )...
%     'measurementOperator', @(x) models.Lorenz1984.customMeasurementOperator(x, avgGaussH) ...
    };
% 
%% 2) Choose method and method parameters
parameters.method = {...
% @()filters.Pajonk.AREnKF()
@()filters.Pajonk.SPCU()
% @()filters.Pajonk.MHF('nsamples', 1000, 'proposalStdDev', 0.2)
% @()filters.Pajonk.KF()
% @()filters.Evensen.SqrtEnKF()
% @()filters.Evensen.EnKF()
% @()filters.Pajonk.AREnKF('screening','gain','transformation_wavelet_enabled', true)
% 'normalization_fourier_enabled',true, ...
% 'normalization_gridPoint_enabled',true ...
%  ...
};

%% 3) Set global parameters and run
parameters.visualize = {true};    % output live graphics?
parameters.progress = {true};     % output progress information to console?

parameters.nRuns = {1};            % number of runs (to create statistics, requires the use of controllers.multiRun)
parameters.statistics = {true};      % compute statistics over the NRUNS or store all runs?
parameters.stat_spectrum = {false};  % compute SVD of ensemble for results? (normally not necessary)
parameters.stat_mean = {true}; % store mean? (this is typically the "best guess")
parameters.stat_relErrors = {true}; % store RMSE/relErr?
parameters.stat_var = {true}; % store variance? (crude measure for the "uncertainty")
parameters.stat_summary = {false}; % store quantiles 0.025, 0.25, 0.5, 0.75, 0.975?
parameters.stat_truth = {true}; % store truth?

% ... and off we go. Interesting here: set the output file name format. You
% can use any input parameter name in {} to have it replaced by the
% respective value. WARNING: only works when the value can be translated
% to a string!
% 
% tools.batchcall(parameters, @controllers.multiRun, 'results/current/test.mat');
results = tools.batchcall(parameters, @controllers.singleRun);