基于遗传算法的机器人路径规划MATLAB源码

基于遗传算法的机器人路径规划MATLAB源码算法的思路如下：取各障碍物顶点连线的中点为路径点，相互连接各路径点，将机器人移动的起点和终点限制在各路径点上，利用Dijkstra算法来求网络图的最短路径，找到从起点P1到终点Pn的最短路径，由于上述算法使用了连接线中点的条件，不是整个规划空间的最优路径，然后利用遗传算法对找到的最短路径各个路径点Pi(i=1,2,…n)调整，让各路径点在相应障碍物端点连线上滑动，利用Pi=Pi1+ti×(Pi2-Pi1)（ti∈[0,1]i=1,2,…n）即可确定相应的Pi，即为新的路径点，连接此路径点为最优路径。function[L1,XY1,L2,XY2]=JQRLJGH(XX,YY)%%基于Dijkstra和遗传算法的机器人路径规划演示程序%GreenSim团队原创作品，转载请注明%GreenSim团队长期从事算法设计、代写程序等业务%欢迎访问GreenSim——算法仿真团队→（包括：边界、障碍物、障碍物顶点之间的连线、Dijkstra的网络图结构）%Fig2由Dijkstra算法得到的最短路径%Fig3由遗传算法得到的最短路径%Fig4遗传算法的收敛曲线（迄今为止找到的最优解、种群平均适应值）%%画Fig1figure(1);PlotGraph;title('地形图及网络拓扑结构')PD=inf*ones(26,26);fori=1:26forj=1:26ifD(i,j)==1x1=XY(i,5);y1=XY(i,6);x2=XY(j,5);y2=XY(j,6);dist=((x1-x2)^2+(y1-y2)^2)^0.5;PD(i,j)=dist;endendend%%调用最短路算法求最短路s=1;%出发点t=26;%目标点[L,R]=ZuiDuanLu(PD,s,t);L1=L(end);XY1=XY(R,5:6);%%绘制由最短路算法得到的最短路径figure(2);PlotGraph;holdonfori=1:(length(R)-1)x1=XY1(i,1);y1=XY1(i,2);x2=XY1(i+1,1);y2=XY1(i+1,2);plot([x1,x2],[y1,y2],'k');holdonendtitle('由Dijkstra算法得到的初始路径')%%使用遗传算法进一步寻找最短路%第一步：变量初始化M=50;%进化代数设置N=20;%种群规模设置Pm=0.3;%变异概率设置LC1=zeros(1,M);LC2=zeros(1,M);Yp=L1;%第二步：随机产生初始种群X1=XY(R,1);Y1=XY(R,2);X2=XY(R,3);Y2=XY(R,4);fori=1:Nfarm{i}=rand(1,aaa);end%以下是进化迭代过程counter=0;%设置迭代计数器whilecounterM%停止条件为达到最大迭代次数%%第三步：交叉%交叉采用双亲双子单点交叉newfarm=cell(1,2*N);%用于存储子代的细胞结构Ser=randperm(N);%两两随机配对的配对表A=farm{Ser(1)};%取出父代AB=farm{Ser(2)};%取出父代BP0=unidrnd(aaa-1);%随机选择交叉点a=[A(:,1:P0),B(:,(P0+1):end)];%产生子代ab=[B(:,1:P0),A(:,(P0+1):end)];%产生子代bnewfarm{2*N-1}=a;%加入子代种群newfarm{2*N}=b;fori=1:(N-1)A=farm{Ser(i)};B=farm{Ser(i+1)};newfarm{2*i}=b;endFARM=[farm,newfarm];%新旧种群合并%%第四步：选择复制SER=randperm(2*N);FITNESS=zeros(1,2*N);fitness=zeros(1,N);fori=1:(2*N)PP=FARM{i};FITNESS(i)=MinFun(PP,X1,X2,Y1,Y2);%调用目标函数endfori=1:Nf1=FITNESS(SER(2*i-1));f2=FITNESS(SER(2*i));iff1=f2elsefarm{i}=FARM{SER(2*i)};fitness(i)=FITNESS(SER(2*i));endend%记录最佳个体和收敛曲线minfitness=min(fitness);meanfitness=mean(fitness);ifminfitnessYppos=find(fitness==minfitness);Xp=farm{pos(1)};Yp=minfitness;endifcounter==10PPP=[0.5,Xp,0.5]';PPPP=1-PPP;X=PPP.*X1+PPPP.*X2;Y=PPP.*Y1+PPPP.*Y2;XY2=[X,Y];figure(3)PlotGraph;holdonfori=1:(length(R)-1)x1=XY2(i,1);y1=XY2(i,2);x2=XY2(i+1,1);y2=XY2(i+1,2);plot([x1,x2],[y1,y2],'k');holdonendtitle('遗传算法第10代')holdonfori=1:(length(R)-1)x1=XY1(i,1);y1=XY1(i,2);x2=XY1(i+1,1);y2=XY1(i+1,2);plot([x1,x2],[y1,y2],'k','LineWidth',1);holdonendendifcounter==20PPP=[0.5,Xp,0.5]';PPPP=1-PPP;X=PPP.*X1+PPPP.*X2;Y=PPP.*Y1+PPPP.*Y2;XY2=[X,Y];figure(4)PlotGraph;holdonfori=1:(length(R)-1)x1=XY2(i,1);y2=XY2(i+1,2);plot([x1,x2],[y1,y2],'k');holdonendtitle('遗传算法第20代')holdonfori=1:(length(R)-1)x1=XY1(i,1);y1=XY1(i,2);x2=XY1(i+1,1);y2=XY1(i+1,2);plot([x1,x2],[y1,y2],'k','LineWidth',1);holdonendendifcounter==30PPP=[0.5,Xp,0.5]';PPPP=1-PPP;X=PPP.*X1+PPPP.*X2;Y=PPP.*Y1+PPPP.*Y2;XY2=[X,Y];figure(5)PlotGraph;holdonfori=1:(length(R)-1)x1=XY2(i,1);y1=XY2(i,2);x2=XY2(i+1,1);y2=XY2(i+1,2);plot([x1,x2],[y1,y2],'k');holdonendtitle('遗传算法第30代')holdonfori=1:(length(R)-1)x1=XY1(i,1);y2=XY1(i+1,2);plot([x1,x2],[y1,y2],'k','LineWidth',1);holdonendendifcounter==40PPP=[0.5,Xp,0.5]';PPPP=1-PPP;X=PPP.*X1+PPPP.*X2;Y=PPP.*Y1+PPPP.*Y2;XY2=[X,Y];figure(6)PlotGraph;holdonfori=1:(length(R)-1)x1=XY2(i,1);y1=XY2(i,2);x2=XY2(i+1,1);y2=XY2(i+1,2);plot([x1,x2],[y1,y2],'k');holdonendtitle('遗传算法第40代')holdonfori=1:(length(R)-1)x1=XY1(i,1);y1=XY1(i,2);x2=XY1(i+1,1);y2=XY1(i+1,2);plot([x1,x2],[y1,y2],'k','LineWidth',1);holdonendendifcounter==50PPP=[0.5,Xp,0.5]';PPPP=1-PPP;X=PPP.*X1+PPPP.*X2;Y=PPP.*Y1+PPPP.*Y2;XY2=[X,Y];figure(7)PlotGraph;holdonfori=1:(length(R)-1)x1=XY2(i,1);y1=XY2(i,2);x2=XY2(i+1,1);y2=XY2(i+1,2);plot([x1,x2],[y1,y2],'k');holdonendtitle('遗传算法第50代')holdonfori=1:(length(R)-1)x1=XY1(i,1);y1=XY1(i,2);x2=XY1(i+1,1);y2=XY1(i+1,2);plot([x1,x2],[y1,y2],'k','LineWidth',1);holdonendendLC2(counter+1)=Yp;LC1(counter+1)=meanfitness;%%第五步：变异fori=1:NifPmrand&&pos(1)~=iAA=farm{i};AA(POS)=rand;farm{i}=AA;endendcounter=counter+1;disp(counter);end%%输出遗传算法的优化结果PPP=[0.5,Xp,0.5]';PPPP=1-PPP;X=PPP.*X1+PPPP.*X2;Y=PPP.*Y1+PPPP.*Y2;XY2=[X,Y];L2=Yp;%%绘制Fig3figure(8)PlotGraph;holdonholdonfori=1:(length(R)-1)x1=XY1(i,1);y1=XY1(i,2);x2=XY1(i+1,1);y2=XY1(i+1,2);plot([x1,x2],[y1,y2],'k','LineWidth',1);holdonendfori=1:(length(R)-1)x1=XY2(i,1);y1=XY2(i,2);x2=XY2(i+1,1);y2=XY2(i+1,2);plot([x1,x2],[y1,y2],'k');holdonendtitle('遗传算法最终结果')figure(9)PlotGraph;holdonfori=1:(length(R)-1)x1=XY1(i,1);y1=XY1(i,2);x2=XY1(i+1,1);y2=XY1(i+1,2);plot([x1,x2],[y1,y2],'k','LineWidth',1);holdonendholdonfori=1:(length(R)-1)x1=XY2(i,1);y1=XY2(i,2);x2=XY2(i+1,1);y2=XY2(i+1,2);plot([x1,x2],[y1,y2],'k','LineWidth',2);holdonendtitle('遗传算法优化前后结果比较')%%绘制Fig4figure(10);plot(LC1);holdonplot(LC2);xlabel('迭代次数');title('收敛曲线');