标签：dap 变异 group 模拟 switch 技术 迭代 江苏省 color
2019-03-10 16:37:59
转眼间就到了第三周,刚考完Python二级,江苏省的题目有点太那个了吧,什么都有,第一道程序题给的最后结果都是错的?
花了我好长时间在那里,一直在找设置精度的函数,然鹅,回到宿舍用matlab计算也是python上跑的结果.呵呵,今天终于开始看算法了
第一个算法就是基因算法.话不多说开始.
基因算法原理就不重述了,直接上代码
程序设计中选择的二进制编码，种群规模二 20，染色体杂交时的概率为 0.25，变异时概 率为 0.01，适应度函数为函数本身，终止条件为达到预定最大迭代次数。
Objfunction & initial group
%n - 繁衍n代 不加上第一代，初始化父代 function gene(n) f = @(x1,x2)21.5+x1*sin(4*pi*x1)+x2*sin(20*pi*x2); % initial group %种群数为20个 V =cell(1,20); for i = 1:20 temp1 = round(rand(1,33)); %33 = 18+15=5+13+2+13 temp2 = vec2str(temp1); V{i} = temp2; end clear temp1 temp2 record = adapt(V,f); maxrec = maxrecord(record,V); %自然选择，保存最适合 tic for i=1:n V = chfather(V,record); V = opcrossover(V); V = variation(V); record = adapt(V,f); temp = maxrecord(record,V); if (temp(1,1)>maxrec(1,1)) maxrec = temp; mark = i; end end time = toc; format long fprintf(‘ x1: %f\n‘,maxrec(2)); fprintf(‘ x2: %f\n‘,maxrec(3)); fprintf(‘f(x1,x2) : %f\n‘,maxrec(1)); fprintf(‘ from: %1.0f(th) generation \n‘,mark); fprintf(‘ time: %f s\n‘,time);
vec2str
%Char2str %将向量转化成字符串 %bin2dec()中的参数为字符串！ function str = vec2str(vec) len = length(vec); str = ‘‘; for i=1:len temp = num2str(vec(i)); str = strcat(str,temp); end end
adapt
function record = adapt(V,f) record = zeros(20,1,‘double‘); for i=1:20 temp1 =V{i}; temp2 = temp1(1:18); temp3 = temp1(19:end); record(i,1) = f(bin_x(temp2,1),bin_x(temp3,2)); end end
maxrecord
function maxrec = maxrecord(record,V) [m,i] = max(record); maxrec(1,1) = m; temp1 = V{i}; temp2 = temp1(1:18); temp3 = temp1(19:end); maxrec(2,1) = bin_x(temp2,1); maxrec(3,1) = bin_x(temp3,2); end
chfather
%选择父代，似乎染色体进行了异位 function V = chfather(V,record) F = sum(record); %pk-平均,qk-累计 pk = record/F; qk = zeros(20,1,‘double‘); for i=1:20 qk(i) = sum(pk(1:i)); end %模拟转盘转动20次 r = rand(20,1); for i = 1:20 k = 1; while(r(i)>qk(k)) k = k+1; end r(i) = k; end temp = V; for i =1:20 V{i} = temp{r(i)}; end end
opcorssover
function V = opcorssover(V) %交换概率？pc？ %one point corssover pc = 0.25; l = 1; while(l==1) r = rand(20,1); mk = find(r<pc); l = numel(mk); end if mod(numel(mk),2) == 1 mk = mk(1:end-1); end ri = randi([1,32],1,numel(mk)/2); for i = 1:numel(mk)/2 [V{mk(2*(i-1)+1)},V{mk(2*(i-1)+2)}] = onecross(V{mk(2*(i-1)+1)},... V{mk(2*(i-1)+2)},ri(i)); end end
variation
function V = variation(V) pm = 0.01; for i = 1:20 r = rand(33,1); k = find(r<pm); for j = 1:numel(k) V{i} = vari(V{i},k(j)); end end end
Vari
function gnew = vari(g_old,pos) gnew = g_old; if gnew(pos) ==‘1‘ gnew(pos) =‘0‘; else gnew(pos) =‘1‘; end end
onecorss
function [g1,g2]=onecross(gen1,gen2,pos) len = numel(gen1); g1 = gen1(1:pos); g2 = gen2(1:pos); g1(pos+1:len) = gen2(pos+1:end); g2(pos+1:len) = gen1(pos+1:end); end
bin_x
function num = bin_x(bin,opt) switch opt case 1 num = -3.0+bin2dec(bin)*((12.1-(-3.0))/(2^18-1)); %12+3.0 = 16 ,0,1,2,3,4 %6-4 = 2 ,0,1,2, %精度为0.5^13 case 2 num = 4.1+bin2dec(bin)*((5.8-4.1)/(2^15-1)); end end
标签：dap 变异 group 模拟 switch 技术 迭代 江苏省 color
原文地址：https://www.cnblogs.com/zero27315/p/10505955.html
