liziqun.rar_粒子群图形_粒子群适应度

clear all clc x=1:1:40; %直接使用路径损耗为4 %求能量，直接将动差函数n=1即可 P_u=1;%用户最大发射功率 % lamda_1=1e-6;%基站密度 % lamda_2=5e-6; lamda_u=5e-5;%用户密度 gama_0=10^(-9);%基站端的接收功率-60dBm alpha=4; sim_num=1; noise_power_per_HZ = 10 ^ (-174 / 10) * 0.001;%噪声功率-174dBm/Hz array_length=length(x); P_1_static=10; P_2_static=100; EE= zeros(array_length , 1); EE_1= zeros(array_length , 1); F_down_and_up_sucess= zeros(array_length , 1); All_success= zeros(array_length , 1); t=0.5;%SINR阈值 lamda_1=1e-6;%基站密度 W=1e6;%带宽 for i=1:1:array_length % t=0.5;%SINR阈值 k=10^(-1+(i-1)*2/(array_length-1));%这是为了表示基站密度关系而加的 lamda_2=5*lamda_1; P_2_max=2; P_1_max=10*P_2_max; M=20;%3*i;%子信道个数 P_2=P_2_max/M; P_1=P_1_max/M;%宏基站功率 delay_threshold_1=0.05; delay_threshold_2=0.001; % theta=1; % t=10^(-2+(i-1)*4/(array_length-1));%SINR阈值 theta=10^((i-1)*4/(array_length-1));%CRE偏置 % for j=1:50 % a=j; % [P_delay_1(j),P_delay_2(j),P_delay(j)]=Delay_probability_invoking(lamda_1,lamda_2,P_1,P_2,alpha,theta,a,lamda_u); % if P_delay_1(j)<delay_threshold_1 & P_delay_2(j)<delay_threshold_2 % M=j; % break % end % end % MMM(i)=20; %wwww=Delay_probability(lamda_1,lamda_2,P_1,P_2,alpha,theta,M,lamda_u); [theta_dB(i),EE(i,1),down_SINR_coverage(i,1),downrate(i,1),uprate(i,1),t_db(i), F_down_and_up_sucess(i,1),F_M_m_1(i,1),y_up_1(i,1),y_down(i,1),k_dB(i),... y_down_1(i,1),y_down_2(i,1),y_up_1_1(i,1),y_up_1_2(i,1),SE(i,1),P_busy_1(i,1),P_busy_2(i,1)... ,y_up_mean_achievable_rate_1(i,1),y_up_mean_achievable_rate_2(i,1),y_down_mean_achievable_rate(i,1),y_down_mean_achievable_rate_1(i,1),... y_down_mean_achievable_rate_2(i,1),Rate_hotspot(i,1),Rate_nonhotspot(i,1),m(i,1),Delay_probability(i,1),EE_up(i,1),SE_up(i,1),EE_all(i,1)]... =DL_Optimal_value_calculate(P_1,P_2,P_1_static,P_2_static,P_u,lamda_1,lamda_2,lamda_u,gama_0,t,alpha,theta,k,M,W); Finish_Precentage = i/(3*array_length) * 100 end gcf1=1; figure(gcf1); hold off; hold on; [AX,H1,H2]=plotyy(theta_dB,EE(:,1),theta_dB,SE(:,1));%'m.-','DisplayName','\lambda_2/\lambda_1=50','LineWidth',1) hold on; set(get(AX(1),'Ylabel'),'String','EE(Mbits/Joule)') %设置Y轴名称 set(get(AX(2),'Ylabel'),'String','SE(bits/s/Hz)') % set(AX(1),'XColor','k','YColor','b');%设置轴颜色 % set(AX(2),'XColor','k','YColor','r'); set(AX(2),'YLim',[5,7]) %Y轴范围 set(AX(2),'YTick',[5:0.1:7]) %Y轴刻度 legend([H1,H2],{'EE';'SE'});%显示图例 % set(H1,'LineStyle','-','color','b'); % set(H2,'LineStyle','-','color','r'); grid on; xlabel('P_1/P_2(dB)') % xlabel('\theta(dB)') gcf2=2; figure(gcf2); hold on; % plot(theta_dB,EE(:,1),'b.-','DisplayName','\lambda_2/\lambda_1=50','LineWidth',1) hold on; plot(theta_dB,y_down_1(:,1),'b.-','DisplayName','Analy:\theta=10dB','LineWidth',1) hold on; plot(theta_dB,y_down_2(:,1),'r.-','DisplayName','Analy:\theta=10dB','LineWidth',1) grid on; xlabel('M') ylabel('SE(Bit/Sec/Hz)') % set(gca,'YLim', [0 0.01], 'XLim',[-60 60] ) gcf3=3; figure(gcf3); % semilogx(m,EE(:,1)) plot(theta_dB,down_SINR_coverage(:,1),'r.-','DisplayName','Analy:\theta=10dB','LineWidth',1) grid on; xlabel('\lambda_2/\lambda_1(dB)') ylabel('h') % % set(gca,'YLim', [0 0.01], 'XLim',[-60 60] ) %% 最优计算 % 参数初始化 % G=1; % popsize = 200; %种群规模 % c1 = 2; % c2 = 2;%学习因子 % gbest(G) = 0; %全局最优解 % best_fitness = zeros(G,1); %全局最优函数值 % %pbest = 0; %局部最优解 % %eita_min = 0; % %eita_max = 1; % itime = 0; %当前迭代次数 % gen = 40; %迭代次数 % max_v = 10; %最大速度 % best_in_history = zeros(G,gen);%初始化全局历史最优解 % % for g=1:G % % pop(popsize,5) = 0; % %初始化种群，第1列为eita值，第2列为速度分量，第3列为个体\eta最优值, % %第4列为个体最优适应值，第5列为当前个体适应值 % % % %1. 初始化粒子位置和速度 % for i=1:popsize % pop(i,1)=10^( rand()*40/10 ); %初始化粒子位置，0-1之间，步长为其速度 % pop(i,3) = pop(i,1); %初始状态个体最优值等于初始位置 % pop(i,2) = 10^((rand()*0.02-0.01)*40/10);%初始化粒子速度，值为-0.01~0.01，间隔为0.0001??????????????????????????????????????粒子速度如何调整合适？ % pop(i,4) = inf; % pop(i,5) = inf; % end % % % gbest(g) = pop(1,1); %全局最优值初始为种群第一个粒子的位置 % % %2. 计算适应值并赋值，寻找局部最优值 % % while itime<=gen % itime = itime + 1; % % for i = 1:popsize % [theta_dB(i),EE(i,1),down_SINR_coverage(i,1),downrate(i,1),uprate(i,1),t_db(i), F_down_and_up_sucess(i,1),F_M_m_1(i,1),y_up_1(i,1),y_down(i,1),k_dB(i),... % y_down_1(i,1),y_down_2(i,1),y_up_1_1(i,1),y_up_1_2(i,1),SE(i,1),P_busy_1(i,1),P_busy_2(i,1)... % ,y_up_mean_achievable_rate_1(i,1),y_up_mean_achievable_rate_2(i,1),y_down_mean_achievable_rate(i,1),y_down_mean_achievable_rate_1(i,1),... % y_down_mean_achievable_rate_2(i,1),Rate_hotspot(i,1),Rate_nonhotspot(i,1),m(i,1),Delay_probability(i,1),EE_up(i,1),SE_up(i,1),EE_all(i,1)]... % =DL_Optimal_value_calculate(P_1,P_2,P_1_static,P_2_static,P_u,lamda_1,lamda_2,lamda_u,gama_0,t,alpha,pop(i,1),k,M,W); % pop(i,5) = EE(i,1); % if pop(i,4)>pop(i,5); %若当前适应值优于个体最优值，则进行更新 % pop(i,4) = pop(i,5); % pop(i,3) = pop(i,1); %位置更新 % end % end % %3.寻找全局最优值 % % if best_fitness(g) < max(pop(:,4)) % best_fitness(g) = max(pop(:,4)); % row = find(pop(:,4) == best_fitness(g)); % gbest(g) = pop(row(1),1); % end % best_in_history(g,itime) = best_fitness(g);%记录当前全局最优值 % % if itime-1>0 % if abs(best_in_history(g,itime-1)-best_in_history(g,itime))==0 % break; % end % end % % %调整粒子速度与位置 % %更新速度 % for i = 1:popsize % pop(i,2) = (0.9+itime./gen.*0.5).*pop(i,2)+c1.*rand().*(pop(i,3)-pop(i,1))+c2.*rand().*(gbest(g)-pop(i,1)); % %pop(i,2) = 0.01*pop(i,2)+c1*rand()*(pop(i,3)-pop(i,1))+c2*rand()*(gbest-pop(i,1)); % if abs(pop(i,2))>max_v % if(pop(i,2)>0) % pop(i,2)=max_v; % else % pop(i,2)=-max_v; % end % end % end % %更新位置 % for i = 1:popsize % pop(i,1)=pop(i,1)+pop(i,2); % end % Finish_Precentage = 2*i/(3*array_length) * 100 % end % gbest_dB(g)=10*log10(gbest(g)); % end % % [Y_col,Ind_row]=min(y_down_mean_achievable_rate_2(:,1))%每列的最大值及行号 [Y_col,Ind_row]=max(SE(:,1))%每列的最大值及行号 % % kk_db=10*log10(theta) % % [Y_col,Ind_row]=max(F_down_and_up_sucess) % %% % gcf3=3; % figure(gcf3); % hold on; % hold on % plot(gbest_dB(1),best_fitness(1),'bd','LineWidth',5,'markersize',8); % gcf4=4; % figure(gcf4); % hold on; % [X,Y]=meshgrid(alpha,t_w); % mesh(alpha,t_w,R_average); % hold on % grid on