Q学习示例--路径搜索matlab完整代码

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %Q_learning of single agent move in N rooms %Matlab Code comption of Qlearning by example %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function q=ReinforcementLearning clc format short; format compact; %two input :R & gamma %immediate reward matrix; %row and column = states; %-inf =no door between room; R = [-inf, -inf , -inf, -inf, 0 , -inf; -inf, -inf , -inf, 0 , -inf , 100; -inf, -inf , -inf, 0 , -inf , -inf; -inf, 0 , 0 ,-inf , 0 , -inf; 0 , -inf , -inf, 0 , -inf , 100; -inf, 0 , -inf, -inf , 0 , 100;]; gamma = 0.8;%learining parameter q = zeros(size(R)); q1 = ones(size(R)); count = 0; for episode = 0:50000 %random initial state y = randperm(size(R,1));%make random date from 1 to 6 a = size(R,1); b = size(R,2); state = y(1); %select any action from this state x = find(R(state,:)>=0);%find possible action of this state if(size(x,1)>0) x1 = RandomPermutation(x); x1 = x1(1); end qMax = max(q,[],2); q(state,x1) = R(state,x1) + gamma*qMax(x1);%get max of all actions %break if convergence: small deviation on q for 1000 consective if sum(sum(abs(q1-q))) < 0.0001 & sum(sum(q>0)) if count>1000 episode; break else count = count +1; end else q1 = q; count = 0; end end %nomalize q g = max(max(q)); if g>0 q = 100*q/g; end