Test_fns.rar_遗传算法

% DEMOGA1.M (DEMO of Genetic Algorithms 1) % % This function displays a number of figures. Select a number of these % figures with the mouse by clicking in the area of the shape. These % figures are the parents for the next generation. The offspring are % created by recombination and mutation. The user is the selector. % % The shape of the figures is defined by Points X- and Y-values each. % These points are ploted by patch(), each figure in a rectangle. % The X- and Y-values are the variables of a individual. % % Aim of this demo is, to "construct" a special figure, e.g. a square % or a star or ..., imagine, what you want and select appropriate points. % % Parameters inside function for changing: % Num - Number of figures (4, 9, 16, 25, ...) % Points - Number of Points per figure (3, 4, 5...), 5 recommended % MAXGEN - Number of generations (more than 30 recommended) % % Syntax: demoga1(); % % Input parameters: % no input parameter % Output parameter: % no output parameter % % Author: Hartmut Pohlheim % History: 24.03.94 file created % 13.01.03 tested under MATLAB v6 by Alex Shenfield function demoga1(); Num = 16; % Number of figures Points = 6; % Number of points per figure MAXGEN = 30; % Number of generations FieldDR = rep([0; 1], [1 Points]); % Fielddescription for mutation % Function needs as many rows as columns, every place has to be filled if sqrt(Num) ~= ceil(sqrt(Num)), error('sqrt(Num) must be an integer'); end ColRowNum = ceil(sqrt(Num)); % Number of columns and rows Shrink=Num/ColRowNum; % Value for shrinking of area SelectNumber = ColRowNum; % Number of parents by selection Cols = 1:1:ColRowNum; % Index of individuals in first columns Rows = 1:ColRowNum:Num; % Index of individuals in first row h = figure; % Open a new figure for output axes('Position',[0 0 1 1]); % Set axes to whole figure set(gca, 'xcolor',[0 0 0]); % Make X-Axis invisible set(gca, 'ycolor',[0 0 0]); % Make Y-Axis invisible axis('ij'); % Position start in left-upper corner axis(axis); % Freeze axes XStart = (Cols(1:ColRowNum-1)/Shrink); % Calculate X-start- and Y-endvalues of lines XLine1 = [XStart; XStart]; % between figures YLine1 = [zeros(1,ColRowNum-1); ones(1,ColRowNum-1)]; % X-end- and Y-startvalues XLine = [XLine1 YLine1]; YLine = [YLine1 XLine1]; % Assemble line vectors MatX=rand(Points,Num); % Create X-values of figures at random MatY=rand(Points,Num); % Create Y-values of figures at random for igen = 1:MAXGEN, % Loop over all generations % Recombine individuals MatXYOff = recombin('recint', [MatX MatY]',NaN, 2); % Mutate individuals MatXYOff = mutbga(MatXYOff, FieldDR, 1/Points); MatXYOff = MatXYOff'; % Invert matrix of individuals MatX = MatXYOff(:,1:Num); % Select X- and Y-values MatY = MatXYOff(:,size(MatXYOff,2)/2+1:size(MatXYOff,2)/2+Num); PolyMatX = MatX / Shrink; % Shrink X-values for fitting in small area PolyMatY = MatY / Shrink; % Shrink Y-values for fitting in small area % Add a value to the X-value for placing individual in the appropriate column for irun = 1:ColRowNum-1, PolyMatX(:,Rows+irun)=PolyMatX(:,Rows+irun)+(irun)/Shrink; end % Add a value to the Y-value for placing individual in the appropriate row for irun = 1:ColRowNum-1, PolyMatY(:,Cols+(irun*ColRowNum))= ... PolyMatY(:,Cols+(irun*ColRowNum))+(irun)/Shrink; end cla; % Clear axes, removes all earlier figures patch(PolyMatX,PolyMatY,'b'); % Plot shape of figures lh = line(XLine, YLine); % Plot lines between figures set(lh,'Color',[.6 .6 .6]); % Set linecolor to grey xclick = []; yclick = []; % Reset vectors for storing click points for isel = 1:SelectNumber, % Loop for as many points as individuals to select set(gcf,'Name',[ ' Select ' int2str(isel) '. figure (of ' ... int2str(SelectNumber) ') with mouseclick (Generation ' ... int2str(igen) ' of ' int2str(MAXGEN) ')']); [xclick1 yclick1] = ginput(1); % get position of mouse click xclick = [xclick; xclick1]; yclick = [yclick; yclick1]; % add position to vectors end xpos = ceil(xclick * Shrink); % Calculate column of mouse click ypos = ceil(yclick * Shrink); % Calculate row of mouse click SelNumber = xpos + ColRowNum * (ypos-1); % Calculate numbers of selected figures % select X- and Y-values of individuals and repeat them MatX = rep(MatX(:,SelNumber),[1 ceil(Num/SelectNumber)]); MatY = rep(MatY(:,SelNumber),[1 ceil(Num/SelectNumber)]); end % End of function