% IN THE NAME OF GOD
% ASBO:A New Optimization Algorithm Based on Average and Subtraction of the Best and Worst Members of the Population for Solving Various Optimization Problems 
% PeerJ 2021
% M. Dehghani,P.Trojovský1, and S. Hubálovský

%% Instructions
% Instructions:
% To use ASBO , you must specify the following:
% 1: The interval of variables (lo, hi) and as well as number of variables (m).
% use the [lo,hi,m]=test_functions_range(F_index);
%
% 2: Define the objective function. objective function or fitness
% use the  fit(i)=test_functions(L,F_index,m);
% L: is the population member
%%
clc;clear;close all;
for F_index=1:23 % objective functions
    display(['PLEASE WAIT, The program is running for F',num2str(F_index)])
    N=50;                                            % Number of population
    [lo,hi,m]=test_functions_range(F_index);
    lo=ones(1,m).*(lo);                              % Lower limit for variables
    hi=ones(1,m).*(hi);                              % Upper limit for variables
    T=1000;                                          % Number of iterations
    fit=[];                                          % objective function vector
    X=[];                                            % population matrix
    X_new=[];                                        % new suggestion population Matrix

    %%
        
        for i=1:m
            X(:,i) = lo(i)+rand(N,1).*(hi(i) -lo(i));              % Initial population
        end
        for i =1:N
            L=X(i,:);
            fit(i)=test_functions(L,F_index,m);                    % Fitness evaluation
        end
        
        for t=1:T  % algorithm iteration
            %% update Best and Worst member (Xb and Xw)
            
            [best , blocation]=min(fit);
            
            if t==1
                X_b=X(blocation,:);                                           % Optimal location
                fbest=best;                                           % The optimization objective function
            elseif best<fbest
                fbest=best;
                X_b=X(blocation,:);
            end
            [worst , wlocation]=max(fit);
            X_w=X(wlocation,:);
            %%
            %%
            L1=0.5*(X_b+X_w); % EQ(3)
            L1=max(L1,lo);L1 = min(L1,hi);
            F1=test_functions(L1,F_index,m);
            L2=(X_b-X_w); % EQ(7)
            L2=max(L2,lo);L2 = min(L2,hi);
            
            %% UPDATE POPULATION
            for i=1:N
                
                %% Update ith population member based on L2 using Eqs(8) and (9)
                I=round(1+rand(1,1));
                X_new(i,:)=X(i,:)+1*rand(1,m) .* L2;
                X_new(i,:) = max(X_new(i,:),lo);X_new(i,:) = min(X_new(i,:),hi);
                
                L=X_new(i,:);
                fit_new(i)=test_functions(L,F_index,m);
                if(fit_new(i)<fit(i))
                    X(i,:) = X_new(i,:);
                    fit(i) = fit_new(i);
                end

                %% Update ith population member based on L1 using Eqs(4)-(6)
                I=round(1+rand(1,m));
                
                if fit(i)> F1
                    X_new(i,:)=X(i,:)+rand(1,m) .* (L1-I.*X(i,:));
                else
                    X_new(i,:)=X(i,:)+rand(1,m) .* (1*X(i,:)-L1);
                end
                
                X_new(i,:) = max(X_new(i,:),lo);X_new(i,:) = min(X_new(i,:),hi);
                
                L=X_new(i,:);
                fit_new(i)=test_functions(L,F_index,m);
                if(fit_new(i)<fit(i))
                    X(i,:) = X_new(i,:);
                    fit(i) = fit_new(i);
                end
                

                %% Update ith population member based on best member using Eqs(10) and (11)
                
                if fit(i)<fbest                                          
                    fbest=fit(i);
                    X_b=X(i,:);
                end
                I=round(1+rand(1,m));
                X_new(i,:)=X(i,:)+rand(1,m) .* (X_b-I.*X(i,:));
                
                X_new(i,:) = max(X_new(i,:),lo);X_new(i,:) = min(X_new(i,:),hi);
                
                L=X_new(i,:);
                fit_new(i)=test_functions(L,F_index,m);
                if(fit_new(i)<fit(i))
                    X(i,:) = X_new(i,:);
                    fit(i) = fit_new(i);
                    
                    %%
                    if fit(i)<fbest                                          
                        fbest=fit(i);
                        X_b=X(i,:);
                    end
                    %%
                end
                %%
            end %% 
            %%
            best_so_far(t)=fbest; % save best solution so far
            average(t) = mean (fit);
        end

    display(['The best optimal value of the objective function found by ASBO is : ', num2str(fbest)]);
    
end