遺傳算法程序代碼--多目標優(yōu)化--函數(shù)最值問題

1、函數(shù)最值問題：F=X2Y2Z2clearclc%%初始化pc=0.9%交叉概率pm=0.05%變異概率popsize=500chromlength1=21chromlength2=23chromlength3=20chromlength=chromlength1chromlength2chromlength3pop=initpop(popsizechromlength)%產(chǎn)生初始種群fi=1:500[objvalue]=calobjva

2、lue(pop)%計算目標函數(shù)值[fitvalue]=calfitvalue(objvalue)%計算個體適應(yīng)度[newpop]=ion(popfitvalue)%選擇[newpop1]=crossover(newpoppc)%交叉[newpop2]=mutation(newpop1pm)%變異[newobjvalue]=newcalobjvalue(newpop2)%計算最新代目標函數(shù)值[newfitvalue]=newcalfitv

3、alue(newobjvalue)%計算新種群適應(yīng)度值[bestindividualbestfit]=best(newpop2newfitvalue)%求出群體中適應(yīng)值最大的個體及其適應(yīng)值y(i)=max(bestfit)%儲存最優(yōu)個體適應(yīng)值pop5=bestindividual%儲存最優(yōu)個體n(i)=i%記錄最優(yōu)代位置%解碼x1(i)=0decodechrom(pop5121)2(pow2(21)1)x2(i)=decodechrom

4、(pop52223)6(pow2(23)1)1x3(i)=decodechrom(pop54520)1(pow2(20)1)pop=newpop2end%%繪圖figure(1)%最優(yōu)點變化趨勢圖i=1:500plot(y(i)b)xlabel(迭代次數(shù))ylabel(最優(yōu)個體適應(yīng)值)title(最優(yōu)點變化趨勢)legend(最優(yōu)點)temp2=decodechrom(pop2223)temp3=decodechrom(pop4520)

5、x1=temp12(pow2(21)1)%將二值域中的數(shù)轉(zhuǎn)化為變量域的數(shù)相當于十進制的Xx2=temp26(pow2(23)1)1x3=temp31(pow2(20)1)objvalue=x1.^2x2.^2x3.^2%計算目標函數(shù)值function[newpop1]=crossover(newpoppc)%交叉%遺傳算法子程序%Name:crossover.m[pxpy]=size(newpop)newpop1=zeros(size(

6、newpop))fi=1:2:px1po=r(1)ifpopccpoint=round(rpy)%隨機尋找交叉點newpop1(i:)=[newpop(i11:cpoint)newpop(icpoint1:py)]%相鄰兩個染色體在交叉點位置交叉newpop1(i1:)=[newpop(i1:cpoint)newpop(i1cpoint1:py)]elsenewpop1(i:)=newpop(i:)%不產(chǎn)生新染色體newpop1(i1: