Optimal Foraging Algorithm最优捕食算法附matlab代码.zip

% % Project Title: Optimal foraging algorithm (OFA) % Developer: Guang-Yu Zhu % Contact Info: zhugy@fzu.edu.cn % % Reference: % [1] Zhu, Guang-Yu, and Zhang, Wei-Bo. "Optimal foraging algorithm for global optimization." Applied Soft Computing 51 (2017): 294-313. %% % This file is the core program of OFA. % Input: Units -> The foraging group with popsize (or N) individuals % The composition of the individual in "Units" is a d+1 dimensional vector, including the d dimensional state vector % and the objective function value % BOUND -> The lower and upper bounds of the state vector % Itermax -> The maximum group's foraging number % Output: GBestIndiv -> The best individual for global % GBestFitness -> The best objective function value of GBestIndiv % BestValProcess -> The change process of GBestFitness function [GBestIndiv,GBestFitness,BestValProcess] = OFA(Units,BOUND,Itermax) xZomeLength = size(Units,2); DIMENSION = xZomeLength-1; % The dimensions of the objective function. d is used for this parameter in Fig.2 of reference [1]. N = size(Units,1); % The number of individuals. [tmpUnitsFitness,uIndex] = sort( Units(:,xZomeLength) ); % Line 7 in Fig.2 of reference [1]. GBestFitness = tmpUnitsFitness(1); GBestIndiv = Units(uIndex(1), : ); clear i j s1 s2 tmpUnitsFitness t = 1; % Line 8 in Fig.2 of reference [1]. BestValProcess(t) = GBestFitness; % keep the change process of GBestFitness %% while( t <= Itermax ) % Line 9 in Fig.2 of reference [1]. clear UnitForaging; K = t/Itermax; for i=1:N % Line 11 in Fig.2 of reference [1]. current = uIndex(i); if (i == 1) % The individuals in the foraging group (Units) has been sorted, so, the first dividual is the best dividual, % Line 12 in Fig.2 for (j=1:DIMENSION) r1=rand(); r2=rand(); UnitForaging(current,j)=Units(current,j)-K*( Units(current,j)-Units(uIndex(N),j) )*r1 + K*( Units(current,j)-Units(uIndex(N),j) )*r2; % Line 15 in Fig.2 end else front=floor(1+(i-1)*rand()); % Line 20 in Fig.2 position = uIndex(front); % Line 20 in Fig.2 for(j=1:DIMENSION) r1=rand(); r2=rand(); UnitForaging(current,j)=Units(current,j)-K*( Units(current,j)-Units(position,j) )*r1+K*( Units(current,j)-Units(position,j) )*r2; % Line 23 in Fig.2 end end for j=1:DIMENSION % Line 16,17 or 24, 25 in Fig.2 if( UnitForaging(current,j)>BOUND(2) ) UnitForaging(current,j) = 2*BOUND(2) - UnitForaging(current,j); end if( UnitForaging(current,j)<BOUND(1) ) UnitForaging(current,j) = 2*BOUND(1) - UnitForaging(current,j); end end % Compute the objective function value of new position. Line 28 in Fig.2 if( Units(current,(1:DIMENSION)) == UnitForaging(current,(1:DIMENSION)) ) UnitForaging(current,xZomeLength) = Units(current,xZomeLength); else [UnitForaging(current,:),UnitForaging(current,xZomeLength)] = Sphere( UnitForaging(current,:) ); end end for i=1:N % Line 30-36 in Fig.2 lambta=rand(); if( (lambta*UnitForaging(i,xZomeLength)/(1+lambta*(t+1))) < (Units(i,xZomeLength)/t) ) Units(i, : )=UnitForaging(i, : ); end end clear r1 r2 s1 s2 lambta [tmpUnitsFitness,uIndex]=sort( Units(:,xZomeLength) ); % Line 37 in Fig.2 if(GBestFitness > tmpUnitsFitness(1)) % Line 38,39 in Fig.2 GBestFitness = tmpUnitsFitness(1); GBestIndiv=Units(uIndex(1), : ); end BestValProcess(t) = GBestFitness; t = t+1; % Line 40 in Fig.2 end