遗传算法执行过程的基本步骤：
- 初始化，确定种群规模，构造染色体。 染色体可以以浮点数或二进制进行构造，即假设求函数最小值，即染色体由函数自变量构成。
- 选择评估函数，计算每个染色体的适应值，确定Best。
- 选择，采用轮盘赌选择算法。
- 交配，确定交配概率，通过生成随机数确定交配对象和交配位。
- 变异，确定变异概率，通过生成随机数确定变异基因。
- 重新评价染色体适应值，确定Best。
- 如果满足算法终止条件，则输出找到的最优解并退出程序，否则返回步骤（3）继续执行。
此处代码是针对 所写的，公式可自由替换。
评价染色体适应值
function [z] = concr(x)
a = x(1);
b = x(2);
c = x(3);
d = x(4);
z = 1/(1+abs(a)+abs(b)+abs(c)+abs(d));
end
用轮盘赌选择交配的染色体
function [y2] = select(y1)
su = sum(y1);
C = ones(1,5);
for j = 1:5
C(j) = y1(j)/su;
end
D = cumsum(C);
E = rand(1,5);
y2 = ones(1,5);
for j = 1:5
for k = 1:4
if E(j)<=D(1)
y2(j) = 1;
elseif E(j)<=D(k+1) && E(j)>=D(k)
y2(j) = k+1;
end
end
end
end
染色体交叉
function [F] = mate(F)
P = rand(1,5);
q = find(P>0.88);
p = length(q);
if p == 2 || p == 3
c1 = F(q(1),:);
c2 = F(q(2),:);
y1 = cross(c1,c2);
F(q(1),:) = y1(:,1:4);
F(q(2),:) = y1(:,5:8);
elseif p == 4 || p ==5
c1 = F(q(1),:);
c2 = F(q(2),:);
y1 = cross(c1,c2);
F(q(1),:) = y1(:,1:4);
F(q(2),:) = y1(:,5:8);
c3 = F(q(3),:);
c4 = F(q(4),:);
y2 = cross(c3,c4);
F(q(3),:) = y2(:,1:4);
F(q(4),:) = y2(:,5:8);
end
end
function y = cross(varargin)
c1 = varargin{1};
c2 = varargin{2};
a = randperm(3);
i = a(1);
b = c1;
switch i
case{1}
c1(:,2:4) = c2(:,2:4);
c2(:,2:4) = b(:,2:4);
case{2}
c1(:,3:4) = c2(:,3:4);
c2(:,3:4) = b(:,3:4);
case{3}
c1(4) = c2(4);
c2(4) = b(4);
end
y = [c1,c2];
end
染色体变异
function [f] = mutate(f)
a = rand(5,4);
[m,n] = find(a<0.1);
b = length(m);
for i = 1:b
f(m(i),n(i)) = randn(1);
end
完整代码
function [ ] = produce( z )
%UNTITLED 此处显示有关此函数的摘要
% 此处显示详细说明
A = randn(5,4); % 随机生成五个染色体
B = ones(1,5);
G = ones(1,z); % 可自行控制迭代次数
for j = 1:z
for i = 1:5
B(i) = concr(A(i,:)); % 评价染色体适应值
end
G(j) = max(B);
num = select(B); % 用旋轮法选择染色体
F = ones(5,4);
for i = 1:5
F(i,:) = A(num(i),:);
end
F = mate(F); % 染色体交叉
F = mutate(F); % 染色体变异
A = F;
end
G
g = 1:1:z;
plot(g,G) % 绘出最优值与代数的图
end
function [y2] = select(y1)
su = sum(y1);
C = ones(1,5);
for j = 1:5
C(j) = y1(j)/su;
end
D = cumsum(C);
E = rand(1,5);
y2 = ones(1,5);
for j = 1:5
for k = 1:4
if E(j)<=D(1)
y2(j) = 1;
elseif E(j)<=D(k+1) && E(j)>=D(k)
y2(j) = k+1;
end
end
end
end
function [z] = concr(x)
a = x(1);
b = x(2);
c = x(3);
d = x(4);
z = 1/(1+abs(a)+abs(b)+abs(c)+abs(d));
end
function [F] = mate(F)
P = rand(1,5);
q = find(P>0.88);
p = length(q);
if p == 2 || p == 3
c1 = F(q(1),:);
c2 = F(q(2),:);
y1 = cross(c1,c2);
F(q(1),:) = y1(:,1:4);
F(q(2),:) = y1(:,5:8);
elseif p == 4 || p ==5
c1 = F(q(1),:);
c2 = F(q(2),:);
y1 = cross(c1,c2);
F(q(1),:) = y1(:,1:4);
F(q(2),:) = y1(:,5:8);
c3 = F(q(3),:);
c4 = F(q(4),:);
y2 = cross(c3,c4);
F(q(3),:) = y2(:,1:4);
F(q(4),:) = y2(:,5:8);
end
end
function y = cross(varargin)
c1 = varargin{1};
c2 = varargin{2};
a = randperm(3);
i = a(1);
b = c1;
switch i
case{1}
c1(:,2:4) = c2(:,2:4);
c2(:,2:4) = b(:,2:4);
case{2}
c1(:,3:4) = c2(:,3:4);
c2(:,3:4) = b(:,3:4);
case{3}
c1(4) = c2(4);
c2(4) = b(4);
end
y = [c1,c2];
end
function [f] = mutate(f)
a = rand(5,4);
[m,n] = find(a<0.1);
b = length(m);
for i = 1:b
f(m(i),n(i)) = randn(1);
end
end