Evolutionary Multi-Objective Optimization Platform

I. Release Notes
The architecture of PlatEMO is simplified, which leads to lower learning cost and higher efficiency. The result file size is also reduced.

  cd(fileparts(mfilename('fullpath')));
    addpath(genpath(cd));
    if isempty(varargin)
        if verLessThan('matlab','8.4')
            errordlg('Fail to establish the GUI of PlatEMO, since the version of MATLAB is lower than 8.4 (R2014b). You can run PlatEMO without GUI by invoking main() with parameters.','Error','modal');
        else
            GUI();
        end
    else
        if verLessThan('matlab','7.14')
            error('Fail to execute PlatEMO, since the version of MATLAB is lower than 7.14 (R2012a). Please update the version of your MATLAB software.');
        else
            Global = GLOBAL(varargin{:});
            Global.Start();
        end
    end
end

2.Higher efficiency
The runtime of Pareto dominance based algorithms is reduced by using a more efficient non-dominated sorting algorithm. The runtime of decomposition based algorithms is reduced due to the new architecture of PlatEMO. The runtime of hypervolume calculation is reduced by new logic and GPU acceleration. In experimental module, the algorithms can be executed in parallel.
3.More conveniences
The populations obtained during the evolutionary process can be saved in result files. The references of each algorithm, problem, operator, and metric are given in the comments of the function. The codes of GUI are now open source.

example

over