% Driving Training-Based Optimization
%%% Designed and Developed by Mohammad Dehghani %%%
function[Best_score,Best_pos,DTBO_curve]=DTBO(SearchAgents,Max_iterations,lowerbound,upperbound,dimension,fitness)
lowerbound=ones(1,dimension).*(lowerbound); % Lower limit for variables
upperbound=ones(1,dimension).*(upperbound); % Upper limit for variables
%%
for i=1:dimension
X(:,i) = lowerbound(i)+rand(SearchAgents,1).*(upperbound(i) - lowerbound(i)); % Initial population
end
for i =1:SearchAgents
L=X(i,:);
fit(i)=fitness(L);
end
%%
for t=1:Max_iterations
%% update the best member
[best , blocation]=min(fit);
if t==1
Xbest=X(blocation,:); % Optimal location
fbest=best; % The optimization objective function
elseif best<fbest
fbest=best;
Xbest=X(blocation,:);
end
%%
XF=[X fit'];
XFsort=sortrows(XF,dimension+1);
X=XFsort(:,1:dimension);
fit=( XFsort(:,1+dimension) )';
N_DI=1+round(0.2*SearchAgents*(1-t/Max_iterations));
DI=XFsort(1:N_DI,1:dimension);
F_DI=( XFsort(1:N_DI,1+dimension) )';
%% update DTBO population
for i=1:SearchAgents
%% Phase 1: training by the driving instructor (exploration)
k_i=randperm(N_DI,1);
DI_ki=DI(k_i);
F_DI_ki=F_DI(k_i);
I=round(1+rand(1,1));
if F_DI_ki< fit (i)
X_P1=X(i,:)+rand(1,1) .* (DI_ki-I.*X(i,:)); % Eq. (5)
else
X_P1=X(i,:)+rand(1,1) .* (1.*X(i,:)-I.*DI_ki); % Eq. (5)
end
X_P1= max(X_P1,lowerbound);X_P1 = min(X_P1,upperbound);
% Update X_i based on Eq(6)
F_P1 = fitness(X_P1);
if F_P1 <= fit (i)
X(i,:) = X_P1;
fit (i)=F_P1;
end
%% END Phase 1: training by the driving instructor (exploration)
%% Phase 2: learner driver patterning from instructor skills (exploration)
PIndex=0.01+0.9*(1-t/Max_iterations);
X_P2=(PIndex).* X(i,:)+(1-PIndex) .* (Xbest); % Eq. (7)
X_P2= max(X_P2,lowerbound);X_P2 = min(X_P2,upperbound);
% Update X_i based on Eq(8)
F_P2 = fitness(X_P2);
if F_P2 <= fit (i)
X(i,:) = X_P2;
fit (i)=F_P2;
end
%% END Phase 2: learner driver patterning from instructor skills (exploration)
%% Phase 3: personal practice (exploitation)
R=0.05;
X_P3= X(i,:)+ (1-2*rand(1,dimension))*R*(1-t/Max_iterations).*X(i,:);% Eq.(9)
X_P3= max(X_P3,lowerbound);X_P3 = min(X_P3,upperbound);
% Update X_i based on Eq(10)
F_P3 = fitness(X_P3);
if F_P3 <= fit (i)
X(i,:) = X_P3;
fit (i)=F_P3;
end
%% END Phase 3: personal practice (exploitation)
end% END for i=1:SearchAgents
best_so_far(t)=fbest;
end% END for t=1:Max_iterations
Best_score=fbest;
Best_pos=Xbest;
DTBO_curve=best_so_far;
end
