%%% Designed and Developed by Pavel Trojovský and Mohammad Dehghani %%%


function[Best_score,Best_pos,EBOA_curve]=EBOA(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


%% INITIALIZATION
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 condidate solution
    [Fbest , Blocation]=min(fit);
    if t==1
        Xbest=X(Blocation,:);                                           % Optimal location
        fbest=Fbest;                                           % The optimization objective function
    elseif Fbest<fbest
        fbest=Fbest;
        Xbest=X(Blocation,:);
    end
    
    [Fworst , wlocation]=max(fit);
    XF=[X fit'];
    XFsort=sortrows(XF,dimension+1);
    X=XFsort(:,1:dimension);
    fit=(XFsort(:,dimension+1))';
    
    %%
    %%  Phase 1: Voting process and holding elections (exploration).
    % UPDATE Ai based EQ(4)
    if Fbest~=Fworst
        A=(fit-Fworst)./(Fbest-Fworst);
    else
        A=ones(SearchAgents,1);
    end
    %
    
    N_C=round(0.1*SearchAgents); % number of candidates
    if N_C<2
        N_C=2;
    end
    CANDID=X(1:N_C,:);
    
    % election holding based EQ(5)
    total_V=zeros(N_C,1);
    for i=N_C+1:SearchAgents
        if A(i)> 0.5*(1+rand)
            total_V(1)=total_V(1)+1;
        else
            k=randperm(N_C-1,1);
            total_V(k+1)=total_V(k+1)+1;
        end
    end
    
    % Counting the votes and appointing the elected leader 
    [a,b]=max(total_V);
    row_no_LEADER=b;
    LEADER= X(b,:);
    F_LEADER=fit(b);
    
    for i=1:SearchAgents
        
        % create new location for X_i using EQ(6)
        I=round(1+rand(1,1));
        if rand>0.5
            RAND=rand(1,dimension);
        else
            RAND=rand;
        end

        if fit(i)> F_LEADER
            X_new_P1(i,:)=X(i,:)+RAND .* (LEADER-I.*X(i,:)); % Eq. (6)
        else
            X_new_P1(i,:)=X(i,:)+RAND .* (X(i,:)-LEADER); % Eq. (6)
        end
        
        X_new_P1(i,:) = max(X_new_P1(i,:),lowerbound);X_new_P1(i,:) = min(X_new_P1(i,:),upperbound);
        
        % update position based on Eq (7)
        L=X_new_P1(i,:);
        fit_new_P1(i)=fitness(L);
        if(fit_new_P1(i)<fit(i))
            X(i,:) = X_new_P1(i,:);
            fit(i) = fit_new_P1(i);
        end
        % END Phase 1: Voting process and holding elections (exploration).
        
        %% Phase 2: Public movement to raise awareness (exploitation).
        % create new location for X_i based Phase2 and EQ(8)
        if rand>0.5
            RAND=rand(1,dimension);
        else
            RAND=rand;
        end
        R=0.02*(1-t/Max_iterations);
        X_new_P2(i,:)= X(i,:)+(-1+2*RAND).*R.*X(i,:);J=randperm(dimension,1);X_new_P2(i,J)=X(i,J)+(lowerbound(J)+rand*(upperbound(J)-lowerbound(J)))./t;
        X_new_P2(i,:) = max(X_new_P2(i,:),lowerbound);X_new_P2(i,:) = min(X_new_P2(i,:),upperbound);
        
        % update position based on Eq (9)
        L=X_new_P2(i,:);
        fit_new_P2(i)=fitness(L);
        if(fit_new_P2(i)<fit(i))
            X(i,:) = X_new_P2(i,:);
            fit(i) = fit_new_P2(i);
        end
        % END PHASE 2: Public movement to raise awareness (exploitation).
        
    end
    %%
    
    best_so_far(t)=fbest;
    average(t) = mean (fit);
    
end
Best_score=fbest;
Best_pos=Xbest;
EBOA_curve=best_so_far;
end