Matlab代码书写规范

标识符命名规则

一般变量

变量命名多采用小驼峰命名方式，如下

var serverForYou

循环变量多采用$i,j,k$等前缀形式命名，应避少用非前缀的命名方式，避免与matlab虚数单位重复

for i_population=1:K

常量、全局变量

常量或者全局变量均采用字母大写形式，如

DECISION

矩阵、结构体、元胞变量

矩阵，结构体及元胞变量多采用各类型英文名+"_"+变量实际代表含义名称，如

cell_mother
struct_population
matrix_fish

结构体属性构建及引用方式

构建

% state: optimization state of one generation
state = struct(...
'currentGen', 1,...         % current generation number
'evaluateCount', 0,...      % number of objective function evaluation
'totalTime', 0,...          % total time from the beginning
'firstFrontCount', 0,...    % individual number of first front
'frontCount', 0,...         % number of front
'avgEvalTime', 0 ...        % average evaluation time of objective function (current generation)
);

引用

options.popSize = 100;                   % populaion size
options.maxGen  = 150;                  % max generation 若此数值小于listbox值max，则不会显示

options.numObj = 2;                     % number of objectives
options.numVar = 2;                     % number of design variables
options.numCons = 2;                    % number of constraints

函数

1. 函数命名多采用小写字母+下划线分割形式命名

non_dominant_sort()

2. get/set前缀多用来访问对象或者属性用
3. compute前缀多用来计算某些量的函数用，如

compute_objectives()

4. is前缀多用来判断

isbool()

代码版式

文件表头

样例

%___________________________________________________________________%
%  Multi-Objective Grey Wolf Optimizer (MOGWO)                      %
%  Source codes demo version 1.0                                    %
%                                                                   %
%  Developed in MATLAB R2011b(7.13)                                 %
%                                                                   %
%  Author and programmer: Seyedali Mirjalili                        %
%                                                                   %
%         e-Mail: ali.mirjalili@gmail.com                           %
%                 seyedali.mirjalili@griffithuni.edu.au             %
%                                                                   %
%       Homepage: http://www.alimirjalili.com                       %
%                                                                   %
%   Main paper:                                                     %
%                                                                   %
%    S. Mirjalili, S. Saremi, S. M. Mirjalili, L. Coelho,           %
%    Multi-objective grey wolf optimizer: A novel algorithm for     %
%    multi-criterion optimization, Expert Systems with Applications,%
%    in press, DOI: http://dx.doi.org/10.1016/j.eswa.2015.10.039    %       %
%                                                                   %
%___________________________________________________________________%

函数说明

样例1：

function [ 函数输出参数] = 函数名( 函数输入)
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%此处为函数名，如果不是函数文件就取消此行
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%创建人：
%日期：
%修改人：
%日期：
%函数变量及功能的简单描述
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
--函数主体

样例2：

function result = nsga2(opt, varargin)
% Function: result = nsga2(opt, varargin)
% Description: The main flowchart of of NSGA-II. Note:
%   All objectives must be minimization. If a objective is maximization, the
%   objective should be multipled by -1.
%
% Syntax:
%   result = nsga2(opt): 'opt' is generated by function nsgaopt().
%   result = nsga2(opt, param): 'param' can be any data type, it will be
%       pass to the objective function objfun().
%
%   Then ,the result structure can be pass to plotnsga to display the
%   population:  plotnsga(result);
%
% Parameters:
%   opt : A structure generated by funciton nsgaopt().
%   varargin : Additional parameter will be pass to the objective functions.
%       It can be any data type. For example, if you call: nsga2(opt, param),
%       then objfun would be called as objfun(x,param), in which, x is the
%       design variables vector.
% Return:
%   result : A structure contains optimization result.
%
%         LSSSSWC, NWPU
%   Revision: 1.2  Data: 2011-07-26
%*************************************************************************

样例3：

%NDSort - Do non-dominated sorting by efficient non-dominated sort.
%
%   FrontNo = NDSort(F,s) does non-dominated sorting on F, where F is the
%   matrix of objective values of a set of individuals, and s is the number
%   of individuals to be sorted at least. FrontNo(i) denotes the front
%   number of the i-th individual. The individuals have not been sorted are
%   assigned a front number of inf.
%
%   FrontNo = NDSort(F,C,s) does non-dominated sorting based on constrained
%   domination, where C is the matrix of constraint values of the
%   individuals. In this case, feasible solutions always dominate
%   infeasible solutions, and one infeasible solution dominates another
%   infeasible solution if the former has a smaller overall constraint
%   violation than the latter.
%
%   In particular, s = 1 indicates finding only the first non-dominated
%   front, s = size(F,1)/2 indicates sorting only half the population
%   (which is often used in the algorithm), and s = inf indicates sorting
%   the whole population.
%
%   [FrontNo,K] = NDSort(...) also returns the maximum front number besides
%   inf.
%
%   Example:
%       [FrontNo,MaxFNo] = NDSort(PopObj,1)
%       [FrontNo,MaxFNo] = NDSort(PopObj,PopCon,inf)

%------------------------------- Reference --------------------------------
% [1] X. Zhang, Y. Tian, R. Cheng, and Y. Jin, An efficient approach to
% nondominated sorting for evolutionary multiobjective optimization, IEEE
% Transactions on Evolutionary Computation, 2015, 19(2): 201-213.
% [2] X. Zhang, Y. Tian, R. Cheng, and Y. Jin, A decision variable
% clustering based evolutionary algorithm for large-scale many-objective
% optimization, IEEE Transactions on Evolutionary Computation, 2018, 22(1):
% 97-112.
%------------------------------- Copyright --------------------------------
% Copyright (c) 2018-2019 BIMK Group. You are free to use the PlatEMO for
% research purposes. All publications which use this platform or any code
% in the platform should acknowledge the use of "PlatEMO" and reference "Ye
% Tian, Ran Cheng, Xingyi Zhang, and Yaochu Jin, PlatEMO: A MATLAB platform
% for evolutionary multi-objective optimization [educational forum], IEEE
% Computational Intelligence Magazine, 2017, 12(4): 73-87".
%--------------------------------------------------------------------------

代码主体

空行

空行起着分割块或段落的作用，块（段落）之间一般留一个或两个空行。

空行使用规则：

逻辑块之间要留空行

%*************************************************************************
% 1. Initialization
%*************************************************************************
numObj = length( pop(1).obj );  % number of objectives

refPoints = opt.refPoints;
refWeight = opt.refWeight;      % weight factor of objectives
if(isempty(refWeight))
    refWeight = ones(1, numObj);
end
epsilon = opt.refEpsilon;
numRefPoint = size(refPoints, 1);

% Determine the normalized factors
bUseFrontMaxMin = false;    % bUseFrontMaxMin : If use the maximum and minimum value in the front as normalized factor.
if( strcmpi(opt.refUseNormDistance, 'ever') )
    % 1) Find possiable (not current population) maximum and minimum value
    %     of each objective.
    obj = vertcat(pop.obj);
    if( ~isfield(opt, 'refObjMax_tmp') )
        opt.refObjMax_tmp = max(obj);
        opt.refObjMin_tmp = min(obj);
    else
        objMax = max(obj);
        objMin = min(obj);
        for i = 1:numObj
            if(opt.refObjMax_tmp(i) < objMax(i))
                opt.refObjMax_tmp(i) = objMax(i);
            end
            if(opt.refObjMin_tmp(i) > objMin(i))
                opt.refObjMin_tmp(i) = objMin(i);
            end
        end
        clear objMax objMin
    end
    objMaxMin = opt.refObjMax_tmp - opt.refObjMin_tmp;
    clear obj
elseif( strcmpi(opt.refUseNormDistance, 'front') )
    % 2) Do not use normalized Euclidean distance.
    bUseFrontMaxMin = true;
elseif( strcmpi(opt.refUseNormDistance, 'no') )
    % 3) Do not use normalized Euclidean distance.
    objMaxMin = ones(1,numObj);
else
    % 3) Error
    error('NSGA2:ParamError', ...
        'No support parameter: options.refUseNormDistance="%s", only "yes" or "no" are supported',...
        opt.refUseNormDistance);
end


%*************************************************************************
% 2. Calculate preference distance pop(:).prefDistance
%*************************************************************************
for fid = 1:length(front)
    % Step1: Calculate the weighted Euclidean distance in each front
    idxFront = front(fid).f;            % idxFront : index of individuals in current front
    numInd = length(idxFront);          % numInd : number of individuals in current front
    popFront = pop(idxFront);           % popFront : individuals in front fid

    objFront = vertcat(popFront.obj);   % objFront : the whole objectives of all individuals

    if(bUseFrontMaxMin)
        objMaxMin = max(objFront) - min(objFront); % objMaxMin : the normalized factor in current front
    end

    % normDistance : weighted normalized Euclidean distance
    normDistance = calcWeightNormDistance(objFront, refPoints, objMaxMin, refWeight);
    
    
    % Step2: Assigned preference distance
    prefDistanceMat = zeros(numInd, numRefPoint);
    for ipt = 1:numRefPoint
        [~,ix] = sort(normDistance(:, ipt));%升序排序，选择距离小的
        prefDistanceMat(ix, ipt) = 1:numInd;%排序位置
    end
    prefDistance = min(prefDistanceMat, [], 2); % 2表示计算每行min 对于参考点距离的rank位置，rank的值当作参考距离
    clear ix

    
    % Step3: Epsilon clearing strategy
    idxRemain = 1:numInd;           % idxRemain : index of individuals which were not processed
    while(~isempty(idxRemain))
        % 1. Select one individual from remains
        objRemain = objFront( idxRemain, :);
        selIdx = randi( [1,length(idxRemain)] );
        selObj = objRemain(selIdx, :);

        % 2. Calc normalized Euclidean distance
        % distanceToSel : normalized Euclidean distance to the selected points
        distanceToSel = calcWeightNormDistance(objRemain, selObj, objMaxMin, refWeight);
        

        % 3. Process the individuals within a epsilon-neighborhood
        idx = find( distanceToSel <= epsilon );     % idx : index in idxRemain
        if(length(idx) == 1)    % the only individual is the selected one
            idxRemain(selIdx)=[];
        else
            for i=1:length(idx)
                if( idx(i)~=selIdx )
                    idInIdxRemain = idx(i);     % idx is the index in idxRemain vector
                    id = idxRemain(idInIdxRemain);
                    
                    % *Increase the preference distance to discourage the individuals
                    % to remain in the selection.
                    prefDistance(id) = prefDistance(id) + round(numInd/2);
                end
            end
            idxRemain(idx) = [];
        end
        
    end

    % Save the preference distance
    for i=1:numInd
        id = idxFront(i);
        pop(id).prefDistance = prefDistance(i);
    end

结构体定义之间要留空行

% pop : current population
% newpop : new population created by genetic algorithm operators
% combinepop = pop + newpop;
pop = repmat( struct(...
    'var', zeros(1,nVar), ...   % 变量值
    'obj', zeros(1,nObj), ...  %目标函数值
    'cons', zeros(1,nCons),...  % 
    'rank', 0,...
    'distance', 0,...
    'prefDistance', 0,...       % preference distance used in R-NSGA-II
    'nViol', 0,...   % The number of constraints 
    'violSum', 0),... % The sum of abs(constraints)
    [1,popsize]);  %struct 默认是行形式存在的，如果[2, popsize]则与数组相同

% state: optimization state of one generation
state = struct(...
'currentGen', 1,...         % current generation number
'evaluateCount', 0,...      % number of objective function evaluation
'totalTime', 0,...          % total time from the beginning
'firstFrontCount', 0,...    % individual number of first front
'frontCount', 0,...         % number of front
'avgEvalTime', 0 ...        % average evaluation time of objective function (current generation)
);

函数定义体之间要留空行

函数体内，完整的控制结构和单独的语句块之间要留空行

GreyWolves=CreateEmptyParticle(GreyWolves_num);

for i=1:GreyWolves_num
    GreyWolves(i).Velocity=0;
    GreyWolves(i).Position=zeros(1,nVar);
    for j=1:nVar
        GreyWolves(i).Position(1,j)=unifrnd(lb(j),ub(j),1);
    end
    GreyWolves(i).Cost=fobj(GreyWolves(i).Position')';
    GreyWolves(i).Best.Position=GreyWolves(i).Position;
    GreyWolves(i).Best.Cost=GreyWolves(i).Cost;
end

GreyWolves=DetermineDomination(GreyWolves);

		% Eq.(3.4) in the paper
        c=2.*rand(1, nVar);
        % Eq.(3.1) in the paper
        D=abs(c.*Beta.Position-GreyWolves(i).Position);
        % Eq.(3.3) in the paper
        A=2.*a.*rand()-a;
        % Eq.(3.9) in the paper
        X2=Beta.Position-A.*abs(D);
        
       
        % Eq.(3.4) in the paper
        c=2.*rand(1, nVar);
        % Eq.(3.1) in the paper
        D=abs(c.*Alpha.Position-GreyWolves(i).Position);
        % Eq.(3.3) in the paper
        A=2.*a.*rand()-a;
        % Eq.(3.10) in the paper
        X3=Alpha.Position-A.*abs(D);

逻辑上密切相关的语句序列之间不要留空行

未完，待续！🌜🌛