【优化算法】校本优化算法（SBO）【含Matlab源码 1432期】

一、获取代码方式

获取代码方式1： 通过订阅紫极神光博客付费专栏，凭支付凭证，私信博主，可获得此代码。

获取代码方式2： 通过紫极神光博客主页开通CSDN会员，凭支付凭证，私信博主，可获得此代码。

获取代码方式3： 完整代码已上传我的资源：【优化算法】校本优化算法（SBO）【含Matlab源码 1432期】 备注：开通CSDN会员，仅只能免费获得1份代码（有效期为开通日起，三天内有效）； 订阅紫极神光博客付费专栏，可免费获得2份代码（有效期为订阅日起，三天内有效）；

二、部分源代码

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')

三、运行结果

四、matlab版本及参考文献

1 matlab版本 2014a

2 参考文献 [1] 包子阳,余继周,杨杉.智能优化算法及其MATLAB实例（第2版）[M].电子工业出版社，2016. [2]张岩,吴水根.MATLAB优化算法源代码[M].清华大学出版社，2017.