Another Particle Swarm Toolbox: psoplotswarm(options,state,flag,ijk) - File Exchange

Code covered by the BSD License

Highlights from
Another Particle Swarm Toolbox

... This is like "soft" boundaries, except that some kind of penalty value
...
... Particle swarm optimization for binary genomes.
...
... Find the minimum of a function using Particle Swarm Optimization.
ackleysfcn(x) Ackley's Function.
binarytestfcn(x)
dejongsfcn(x)
dropwavefcn(x)
evolutioncomplete(options,sta... Plays a notification when genetic algorithm finishes. Requires a
griewangksfcn(x) Template for writing custom test/demonstration functions for psodemo.
langermannsfcn(x)
nonlinearconstrdemo(x) Nonlinear constraints demo with Rosenbrock's function.
psoboundsabsorb(state,Aineq,b...
psodemo(DemoMode) Runs the PSO on a few demonstration functions, which should be located
psoiterate(options,state,flag... Updates swarm positions and velocities. Called to iterate the swarm from
psooptimset(varargin) Creates an options structure for pso.
psoplotbestf(options,state,fl... Plots the best, mean, and worst scores of particle swarm.
psoplotscores(options,state,f... Plots the best and mean scores of particle swarm.
psoplotswarm(options,state,fl... Plots the positions of particle swarm.
psoplotswarmsurf(options,stat... Shows the evolution of a 2-variable population over a known surface. This
rastriginsfcn(x) Takes row inputs. If not row, will attempt to correct it.
rosenbrocksfcn(x) Takes row inputs. If input is not row, will attempt to correct it.
schwefelsfcn(x) Schwefel's function.
templatefcn(x) Template for writing custom test/demonstration functions for psodemo.
testfcn1(x)
View all files

from Another Particle Swarm Toolbox by Sam
Implementation of a PSO algorithm with the same syntax as the Genetic Algorithm Toolbox.

psoplotswarm(options,state,flag,ijk)

function state = psoplotswarm(options,state,flag,ijk)
% Plots the positions of particle swarm.
%
% ijk is an additional parameter, in the form of a 1x3 vector. It lists the
% dimension of the problem to be plotted, for a multidimensional problem.
%
% Updated this to conform with MATLAB GA toolbox specifications.

% We'll only plot the first two dimensions by default. User can
% set different dimensions to be plotted using ijk.

if nargin < 4 && size(state.Population,2) > 1 % Go to defaults
    ijk = [1,2] ; % k doesn't exist
elseif size(state.Population,2) == 1 % For 1-D problems
    ijk = 1 ;
else % Robust input check
    ijk = reshape(ijk,1,[]) ;
end

% Input checking
if size(ijk,2) > 3
    warning('PSO:Plotting:tooManyDimensions',...
        'Unable to plot %g dimensions. Plotting first 3 only',...
        length(ijk))
end % if length

if strcmpi(flag(1:4),'init') % Initialize
    delete(findobj(gca,'-regexp','Tag','*Locations'))
    initLoc = line(state.Population(:,ijk(1)),state.Population(:,ijk(2)),...
        'Color',0.75*ones(1,3),...
        'Marker','.',...
        'LineStyle','none',...
        'Tag','Initial Locations') ;
    
    % Set reasonable axes limits
    % ---------------------------------------------------------------------
    xlim([options.PopInitRange(1,ijk(1)) options.PopInitRange(2,ijk(1))])
    if size(ijk,2) > 1
        ylim([options.PopInitRange(1,ijk(2)) ...
            options.PopInitRange(2,ijk(2))])
        if size(ijk,2) > 2
            zlim([options.PopInitRange(1,ijk(3)) ...
                options.PopInitRange(2,ijk(3))])
        end % if size
    end % if size
    % ---------------------------------------------------------------------
    
    title('Swarm positions')
    set(gca,'Tag','Swarm Plot','NextPlot','add')
    
    % Initialize plots
    % ---------------------------------------------------------------------
    if size(ijk,2) == 1 %  One dimensional
        currentLoc = line(state.Population(:,ijk(1)),...
            zeros(size(state.Population,1),1)) ;
    elseif size(ijk,2) == 2 % Two dimensional
        currentLoc = line(state.Population(:,ijk(1)),...
            state.Population(:,ijk(2))) ;
    elseif size(ijk,2) == 3 % Three dimensional
        currentLoc = line(state.Population(:,ijk(1)),...
            state.Population(:,ijk(2)),...
            state.Population(:,ijk(3))) ;
    end % if size
    set(currentLoc,...
            'LineStyle','none',...
            'Marker','.',...
            'Color','blue',...
            'Tag','Swarm Locations') ;
    % ---------------------------------------------------------------------
elseif strcmpi(flag(1:4),'iter') % Iterate
    currentLoc = findobj(gca,'Tag','Swarm Locations','Type','line') ;
    if size(ijk,2) == 1 %  One dimensional
        set(currentLoc,'XData',state.Population(:,ijk(1)))
    elseif size(ijk,2) == 2 % Two dimensional
        set(currentLoc,...
            'XData',state.Population(:,ijk(1)),...
            'YData',state.Population(:,ijk(2)))
    elseif size(ijk,2) == 3 % Three dimensional
        set(currentLoc,...
            'XData',state.Population(:,ijk(1)),...
            'YData',state.Population(:,ijk(2)),...
            'ZData',state.Population(:,ijk(3)))
    end
end

if strcmpi(flag(1:4),'init')
    legend([initLoc,currentLoc],{'Initial Positions','Current Positions'})
end

Highlights from Another Particle Swarm Toolbox

Highlights from
Another Particle Swarm Toolbox