function SBO
%% This function implements the basic School Based Optimization (SBO) algorithm for 10-bar truss optimization
% For more information about this method and other algorithms check the following papers:
%%
global D
% Specity SBO parameters
Itmax=300; % Maximum number of iterations
NClass=5; % Number of classes in the school
PopSize=15; % Population size of each class
% Optimization problem parameters
D=Data10; % For truss function evaluate the functio to get the initial parameters
LB=D.LB; % Lowerbound
UB=D.UB; % Upperbound
FN='ST10'; % Name of analyzer function
%% Randomely generate initial designs between LB and UB
Cycle=1;
for I=1:PopSize
for NC=1:NClass
Designs{NC}(I,:)=LB+rand(1,size(LB,2)).*(UB-LB); % Row vector
end
end
% Analysis the designs
for NC=1:NClass
[PObj{NC},Obj{NC}]=Analyser(Designs{NC},FN);
Best{NC}=[];
end
%% SBO loop
for Cycle=2:Itmax
for NC=1:NClass
% Identify best designs and keep them
[Best{NC},Designs{NC},PObj{NC},Obj{NC},WMeanPos{NC}]=Specifier(PObj{NC},Obj{NC},Designs{NC},Best{NC});
TeachersPObj(NC,1)=Best{NC}.GBest.PObj;
TeachersDes(NC,:)=Best{NC}.GBest.Design;
end
for NC=1:NClass
% Select a teacher
SelectedTeacher=TeacherSelector(Best,NC,TeachersPObj);
% Apply Teaching
[Designs{NC},PObj{NC},Obj{NC}]=Teaching(LB,UB,Designs{NC},PObj{NC},Obj{NC},TeachersDes(SelectedTeacher,:),WMeanPos{NC},FN);
[Best{NC},Designs{NC},PObj{NC},Obj{NC},WMeanPos{NC}]=Specifier(PObj{NC},Obj{NC},Designs{NC},Best{NC});
% Apply Learning
[Designs{NC},PObj{NC},Obj{NC}]=Learning(LB,UB,Designs{NC},Obj{NC},PObj{NC},FN);
[Best{NC},Designs{NC},PObj{NC},Obj{NC},WMeanPos{NC}]=Specifier(PObj{NC},Obj{NC},Designs{NC},Best{NC});
end
% Find best so far solution and Mean
CumPObj=[];
for NC=1:NClass
ClassBestPObj(NC,1)=Best{NC}.GBest.PObj;
ClassMean(NC,1)=mean(PObj{NC});
CumPObj=[CumPObj;PObj{NC}];
end
[~,b]=min(ClassBestPObj);
OveralBestPObj=Best{b}.GBest.PObj;
OveralBestObj=Best{b}.GBest.Obj;
OveralBestDes=Best{b}.GBest.Design;
% Plot time history of the best solution vs. iteration and print the
% results
hold on;plot(Cycle,Best{b}.GBest.PObj,'b*');xlabel('Iteration');ylabel('Best solution value');pause(0.0001)
fprintf('Cycle: %6d, Best (Penalized): %6.4f, Objective: %6.4f\n',Cycle,OveralBestPObj,OveralBestObj);
end
Solution.PObj=OveralBestPObj;% Objective value for best non-penalized solution
Solution.Design=OveralBestDes;% Design for best non-penalized solution
%% Save the results
save('SBO_Results.mat','Solution')
