SOM Neural Network Based Method for Task Assignment of Multi-Age...

%% Task Assignment Algorithm Based on SOM Neural Network Method % 作者：李欣 % 单位：上海海事大学水下机器人与智能系统实验室 % Date: 2009-11-30 % Modified: 2010-1-5, 2014-12-1, 2018-3-5 close all; % 设置搜索范围 xmin = [0; 0]; xmax = [50;50]; % 设定迭代次数 Step = 15; % 设置AUV和目标点的参数 Dmin = 1; count = 0; Target=[15 15; 20 20; 10 10; 5 7; 11 17; 22 9 ]'; Robot = [6 20 11 16 18 19 20 10 8 ;6 16 17 14 11.9 19 8 16 9]; Mat = size(Target); TarNum = Mat(1,2); figure, plot(Robot(1,:),Robot(2,:),'.r','MarkerSize',10); axis([0 30 0 30]); hold on plot(Target(1,:),Target(2,:),'.g','MarkerSize',50); hold off Weight{1} = Robot; Target_a = Target; B = 0.2; % 可调参数 r = 0.5; % 可调参数 w = 1; temp = -1; tt = 1; % 主循环 for t = 1:Step for ii=1:TarNum % 计算距离并排序 [Dist{t}(ii),ind(ii),distn] = somdist(Target_a(:,ii),Weight{w}); Nk = ii; Nm = ind(Nk); winnerR = Weight{w}(:,Nm); winnerT = Target_a(:,Nk); if(Dist{t}(ii) <= 0.1) temp(tt) = ii; tt = tt+1; end flag = 1; for zz = 1:tt-1 if(ii == temp(zz)) flag = 0; end end if(flag == 1) % 获胜神经元邻域更新 Neigh{w} = NeighborFunction(Weight{w},winnerR,r,t); Weight{w+1} = updateWeight(Weight{w},B,Neigh{w},distn,winnerT,Dist{t}(ii)); w = w + 1; end end count = count +1; if(count == 1) figure; plot(Target(1,:),Target(2,:),'.g','MarkerSize',50); hold on plot(Weight{w}(1,:),Weight{w}(2,:),'.r','MarkerSize',10); axis([0 30 0 30]); hold off count = 0; pause(0.2) end end