1 简介
2 部分代码
clc;
clear;
close all;
model =CreateModel();
tic;
plotmap(model);
global_chromosome =Muti_Uav_Ga(model);
toc;
function [ cost,sol,costs ] = FitnessFunction( chromosome,model )
%UNTITLED2 Summary of this function goes here改进:用局部极值来进行交叉操作
% Detailed explanation goes here
for uav=1:model.UAV
x= zeros(1,model.dim);
y= zeros(1,model.dim);
z = zeros(1,model.dim);
%取第uav个航路的坐标
for i=1:model.dim
x(i) = chromosome.pos(i,1,uav);
y(i) = chromosome.pos(i,2,uav);
z(i) = chromosome.pos(i,3,uav);
end
sx = model.sx(uav);
sy = model.sy(uav);
sz = model.sz(uav);
ex = model.endp(1);
ey =model.endp(2);
ez=model.endp(3);
xobs = model.xobs;
yobs = model.yobs;
zobs = model.zobs;
robs = model.robs;
XS=[sx x ex];
YS=[sy y ey];
ZS=[sz z ez];
k =numel(XS);
TP =linspace(0,1,k);
tt =linspace(0,1,50);
xx =[];
yy =[];
zz=[];
for i=1:k-1
%每一段向量分成10个点
x_r = linspace(XS(i),XS(i+1),10);
y_r= linspace(YS(i),YS(i+1),10);
z_r =linspace(ZS(i),ZS(i+1),10);
xx = [xx,x_r];
yy = [yy,y_r];
zz =[zz ,z_r];
end
%calc L
dx =diff(xx);
dy =diff(yy);
dz = diff(zz);
Length = sum(sqrt(dx.^2+dy.^2+dz.^2));
nobs = numel(xobs);
violation=0;
for i=1:nobs
d = sqrt( (xx-xobs(i)).^2+(yy-yobs(i)).^2 );
v = max(1-d/robs(i),0);
violation = violation + mean(v);
end
sol(uav).TP=TP;
sol(uav).XS =XS;
sol(uav).YS=YS;
sol(uav).ZS=ZS;
sol(uav).tt=tt;
sol(uav).xx=xx;
sol(uav).yy=yy;
sol(uav).zz=zz;
sol(uav).dx=dx;
sol(uav).dy=dy;
sol(uav).dz=dz;
sol(uav).Length=Length;
sol(uav).violation=violation;
sol(uav).IsFeasible=(violation==0);
%计算协调适应值
% 3、飞行高度限制
high=0;
for k=1:numel(XS)
x=XS(k);
y=YS(k);
h=terrain(x,y);
if ZS(k)<=(h+10) %限制飞行最低高度
high=high+10000;
elseif ZS(k)>375 %限制飞行最高高度
high=high+10000;
else
high=high+abs(ZS(k)-287); %计算与理想高度差距和
end
end
%z
w1 =0.03;
w2=0.3;
w3=0.1;
w4=0.2;
%markov evaluatea
%获取所有维度的坐标
r_xx=[];r_yy=[];r_zz=[];
for i=2:numel(XS)-2
%每一段向量分成10个点
r_x = linspace(XS(i),XS(i+1),3);
r_y= linspace(YS(i),YS(i+1),3);
r_z =linspace(ZS(i),ZS(i+1),3);
r_xx = [r_xx,r_x];
r_yy = [r_yy,r_y];
r_zz =[r_zz ,r_z];
end
Allpos = [r_xx',r_yy',r_zz'];
[stateProbabilityProcess, expectedCostProcess]=MarkovEvaluate(Allpos,model);
sol(uav).MarkovState = stateProbabilityProcess;
sol(uav).MarkovCost = expectedCostProcess;
sol(uav).costs=[w1*Length,w3*high,w4*mean(expectedCostProcess)*1000];
single_cost(uav)= w1*Length +w3*high+w4*mean(expectedCostProcess)*1000;
sol(uav).cost =single_cost(uav);
end
cost =sum(single_cost);
costs=single_cost;
end
3 仿真结果
4 参考文献
[1]江冰, 郭彭. 基于粒子群算法的三维无人机路径规划方法及规划系统:.
原创文章,作者:carmelaweatherly,如若转载,请注明出处:https://blog.ytso.com/175072.html