【智能优化算法】矮猫鼬优化算法(Dwarf Mongoose Optimization Algorithm,DMHO)

矮猫鼬优化算法(Dwarf Mongoose Optimization Algorithm,DMHO)是期刊“COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING”（IF 7.3）的2022年智能优化算法

01.引言

矮猫鼬优化算法(Dwarf Mongoose Optimization Algorithm,DMHO)模仿矮猫鼬的觅食行为。猫鼬捕食的限制性模式极大地影响了猫鼬的社会行为和生态适应，以补偿高效的家庭营养。猫鼬的代偿性行为适应包括猎物大小、空间利用、群体大小和食物供应。提出的算法中使用了矮猫鼬的三个社会群体:阿尔法群体、保姆群体和侦察兵群体。整个家族以觅食为单位，雌性首领发起觅食，决定觅食路径、覆盖的距离和睡觉的土丘。一定数量的猫鼬(通常是雄性和雌性的混合)充当保姆。它们一直陪伴着幼崽，直到企鹅群中午或傍晚返回。保姆被交换为第一个与群体一起觅食的人(开发阶段)。矮猫鼬不为幼崽筑巢;它们把它们从一个熟睡的土堆移到另一个，不会回到之前觅食的地方。矮猫鼬在一块足以支持整个群体的领土上采用了半游牧的生活方式(探索阶段)。这种游牧行为防止了对某一特定地区的过度开发。它还确保了对整个领土的探索，因为以前参观过的沉睡的土丘不会被归还。

02.优化算法的流程

03.论文中算法对比图

04.部分代码

function [BEF,BEP,BestCost]=DMOA(nPop,MaxIt,VarMin,VarMax,nVar,F_obj)
%nVar=5;             % Number of Decision Variables
VarSize=[1 nVar];   % Decision Variables Matrix Size
%VarMin=-10;         % Decision Variables Lower Bound
%VarMax= 10;         % Decision Variables Upper Bound
%% ABC Settings
% MaxIt=1000;              % Maximum Number of Iterations
% nPop=100;               % Population Size (Family Size)
nBabysitter= 3;         % Number of babysitters
nAlphaGroup=nPop-nBabysitter;         % Number of Alpha group
nScout=nAlphaGroup;         % Number of Scouts
L=round(0.6*nVar*nBabysitter); % Babysitter Exchange Parameter 
peep=2;             % Alpha female痴 vocalization 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Empty Mongoose Structure
empty_mongoose.Position=[];
empty_mongoose.Cost=[];
% Initialize Population Array
pop=repmat(empty_mongoose,nAlphaGroup,1);
% Initialize Best Solution Ever Found
BestSol.Cost=inf;
tau=inf;
Iter=1;
sm=inf(nAlphaGroup,1);
% Create Initial Population
for i=1:nAlphaGrouppop(i).Position=unifrnd(VarMin,VarMax,VarSize);pop(i).Cost=F_obj(pop(i).Position);if pop(i).Cost<=BestSol.CostBestSol=pop(i);end
end
% Abandonment Counter
C=zeros(nAlphaGroup,1);
CF=(1-Iter/MaxIt)^(2*Iter/MaxIt);
% Array to Hold Best Cost Values
BestCost=zeros(MaxIt,1);
%% DMOA Main Loop
for it=1:MaxIt% Alpha groupF=zeros(nAlphaGroup,1);MeanCost = mean([pop.Cost]);for i=1:nAlphaGroup% Calculate Fitness Values and Selection of AlphaF(i) = exp(-pop(i).Cost/MeanCost); % Convert Cost to FitnessendP=F/sum(F);% Foraging led by Alpha femalefor m=1:nAlphaGroup% Select Alpha femalei=RouletteWheelSelection(P);% Choose k randomly, not equal to AlphaK=[1:i-1 i+1:nAlphaGroup];k=K(randi([1 numel(K)]));% Define Vocalization Coeff.phi=(peep/2)*unifrnd(-1,+1,VarSize);% New Mongoose Positionnewpop.Position=pop(i).Position+phi.*(pop(i).Position-pop(k).Position);% Evaluationnewpop.Cost=F_obj(newpop.Position);% Comparisionif newpop.Cost<=pop(i).Costpop(i)=newpop;elseC(i)=C(i)+1;endend   % Scout groupfor i=1:nScout% Choose k randomly, not equal to iK=[1:i-1 i+1:nAlphaGroup];k=K(randi([1 numel(K)]));% Define Vocalization Coeff.phi=(peep/2)*unifrnd(-1,+1,VarSize);% New Mongoose Positionnewpop.Position=pop(i).Position+phi.*(pop(i).Position-pop(k).Position);% Evaluationnewpop.Cost=F_obj(newpop.Position);% Sleeping mouldsm(i)=(newpop.Cost-pop(i).Cost)/max(newpop.Cost,pop(i).Cost);% Comparisionif newpop.Cost<=pop(i).Costpop(i)=newpop;elseC(i)=C(i)+1;endend    % Babysittersfor i=1:nBabysitterif C(i)>=Lpop(i).Position=unifrnd(VarMin,VarMax,VarSize);pop(i).Cost=F_obj(pop(i).Position);C(i)=0;endend    % Update Best Solution Ever Foundfor i=1:nAlphaGroupif pop(i).Cost<=BestSol.CostBestSol=pop(i);endend    % Next Mongoose Positionnewtau=mean(sm);for i=1:nScoutM=(pop(i).Position.*sm(i))/pop(i).Position;if newtau>taunewpop.Position=pop(i).Position-CF*phi*rand.*(pop(i).Position-M);elsenewpop.Position=pop(i).Position+CF*phi*rand.*(pop(i).Position-M);endtau=newtau;end% Update Best Solution Ever Foundfor i=1:nAlphaGroupif pop(i).Cost<=BestSol.CostBestSol=pop(i);endend% Store Best Cost Ever FoundBestCost(it)=BestSol.Cost;BEF=BestSol.Cost;BEP=BestSol.Position;% Display Iteration Informationdisp(['Iteration ' num2str(it) ': Best Cost = ' num2str(BestCost(it))]);end
end
function i=RouletteWheelSelection(P)r=rand;C=cumsum(P);i=find(r<=C,1,'first');
end