matlab7.0版本中遗传算法怎么设置变量范围呢
求助高手：matlab 7.0中自带的遗传算法工具箱怎么在里面设置变量范围呢？它的gatool里面只有initial range啊？谢谢！
[ 本帖最后由 eight 于 2007-10-6 22:28 编辑 ] 我也遇到了此问题，翻了很久的资料和书（包括上网），也没有发现答案。高手帮忙，谢谢啊
回复 沙发 dunmin 的帖子
这样就是定义了（0.1，1）（0.3，3）这样的两个变量；但是我现在也遇到了一个问题，就是定了之后，只有第一个种群数在里面，后面的都不在里面了，而且还是负数，不知道是不是我交叉变异那里是否出了问题！
迷惑不解
有谁知道matlab7.0的变量范围定义问题。我也遇到了同样问题，随后的种群不在最初定义的范围，也是到最后得到的适应度是负数。:@L 只需查看option中的变异函数mutationgaussian.m 然后打开它：open mutationgaussian 。在71行 mutationChildren(i,:) = parent+ scale .* randn(1,length(parent));后面加上两行来控制范围即可。
mutationChildren(i,find(mutationChildren(i,:)>upper))=upper(find(mutationChildren(i,:)>upper));
mutationChildren(i,find(mutationChildren(i,:)<lower))=lower(find(mutationChildren(i,:)<lower));
整个函数mutationgaussian 的内容为：
function mutationChildren = mutationgaussian(parents,options,GenomeLength,FitnessFcn,state,thisScore,thisPopulation,scale,shrink)
%MUTATIONGAUSSIAN Gaussian mutation.
% MUTATIONCHILDREN = MUTATIONGAUSSIAN(PARENTS,OPTIONS,GENOMELENGTH,...
% FITNESSFCN,STATE,THISSCORE,THISPOPULATION,SCALE, ...
% SRHINK) Creates the mutated children using gaussian
% mutation. Mutated genes are gaussianly distributed about their parents
% values.
%
% SCALE controls what fraction of the gene's range is searched. A
% value of 0 will result in no change, a SCALE of 1 will result in a
% distribution whose standard deviation is equal to the range of this gene.
% Intermediate values will produce ranges in between these extremes.
%
% SHRINK controls how fast the SCALE is reduced as generations go by.
% A SHRINK value of 0 will result in no shrinkage, yielding a constant search
% size. A value of 1 will result in SCALE shrinking linearly to 0 as
% GA progresses to the number of generations specified by the options
% structure. (See 'Generations' in GAOPTIMSET for more details). Intermediate
% values of SHRINK will produce shrinkage between these extremes.
% Note: SHRINK may be outside the interval (0,1), but this is ill-advised.
%
% Example:
% options = gaoptimset('MutationFcn',{@mutationgaussian});
%
% This specifies that the mutation function used will be
% MUTATIONGAUSSIAN, and since no values for SCALE or SHRINK are specified
% the default values are used.
%
% scale = 0.5; shrink = 0.75;
% options = gaoptimset('MutationFcn',{@mutationgaussian,scale,shrink});
%
% This specifies that the mutation function used will be
% MUTATIONGAUSSIAN, and the values for SCALE or SHRINK are specified
% as 0.5 and 0.75 respectively.
%
% Copyright 2003-2004 The MathWorks, Inc.
% $Revision: 1.10.4.1 $$Date: 2004/08/20 19:48:37 $
% Use default parameters if the are not passed in.
% If these defaults are not what you prefer, you can pass in your own
% values when you set the mutation function:
%
% options.MutationFunction = { mutationgaussian, 0.3, 0} ;
%
if(strcmpi(options.PopulationType,'doubleVector'))
if(nargin < 9)
shrink = 1;
if(nargin < 8)
scale = 1;
end
end
if (shrink > 1) || (shrink < 0)
msg = sprintf('Shrink factors that are less than zero or greater than one may \n\t\t result in unexpected behavior.');
warning('gads:MUTATIONGAUSSIAN:ShrinkFactor',msg);
end
scale = scale - shrink * scale * state.Generation/options.Generations;
range = options.PopInitRange;
lower = range(1,:);
upper = range(2,:);
scale = scale * (upper - lower);
mutationChildren = zeros(length(parents),GenomeLength);
for i=1:length(parents)
parent = thisPopulation(parents(i),:);
mutationChildren(i,:) = parent+ scale .* randn(1,length(parent));
mutationChildren(i,find(mutationChildren(i,:)>upper))=upper(find(mutationChildren(i,:)>upper));
mutationChildren(i,find(mutationChildren(i,:)<lower))=lower(find(mutationChildren(i,:)<lower));
end
elseif(strcmpi(options.PopulationType,'bitString'))
% there's no such thing as binary gaussian mutation se we'll just
% revert to uniform.
mutationChildren = mutationuniform(parents ,options, GenomeLength,FitnessFcn,state, thisScore,thisPopulation);
end
[ 本帖最后由 ChaChing 于 2009-2-16 08:20 编辑 ] 谢谢楼上的解答，我也遇到了这个问题，照你说的改改试试
页:
[1]