寄生捕食算法.zip

% 寄生捕食算法 % % function [fmin,bestnest,Convergence_curve]=PPA(n,maxiter,lb,ub,d,fobj) fitness=inf*ones(n,1); if length(lb) < d lb=lb*ones(1,d);ub=ub*ones(1,d); end % Random initial solutions using Eq.(5) nest =init(n,ub,lb); [nest,fitness]=get_best_nest(fobj,nest,nest,fitness); % Get the current best [fitness,I]=sort(fitness);nest=nest(I,:); fmin=fitness(1); bestnest=nest(1,:); Convergence_curve(1)=fmin; wMax = 0.9; wMin = 0.3; w = linspace(wMax, wMin, maxiter); Cats_v = 0.25 * nest; % Initial velocity of tracking mode cats % Growth rate (Section 3.1.2.1, Fig. 3) [GrowthRateCrows,GrowthRateCats,GrowthRateCuckoos] =Growth_rate(n,maxiter); % Starting iterations for t=2:maxiter nCats=GrowthRateCats(t);%current no. of Cats nCrows=GrowthRateCrows(t);%current no. of Crows nCuckoos=GrowthRateCuckoos(t);%current no. of Cuckoos % __________________ Nesting Phase: Crows _____________________________ % The current best group discover Crow's nest using Eqs.(6-9) new_Crows=get_Crows(nest(1:nCrows,:),lb,ub); % _______________ crow-cuckoo subsystem: Parasitism phase _________ % Select parasitized nests based on their fitness Rolette_index = RankingSelection(n, nCuckoos); parasitized_nests = nest(Rolette_index(1:nCuckoos),:); % Replace some Crow's eggs (host) by Cuckoo's eggs % (the better nests the higher chance for parasitism) new_Cuckoos = get_Cuckoos(parasitized_nests,lb,ub,t,maxiter); % Choosing non-parasitized nests, randomly non_parasitized_nests = randperm( n); % cuckoo chicks reveal repulsive compounds that repel cats. % Eliminate nests with high malodorous secretion that occupied by Cuckoos non_parasitized_nests(Rolette_index(1:nCuckoos)) = []; % Fixing % Select nests with low malodorous secretion Cats_nests = nest(non_parasitized_nests( 1:nCats),:); % Choosing new_Cats_v = Cats_v(non_parasitized_nests( 1:nCats),:); % Cats track the nest with a low malodorous secretion [new_Cats, new_Cats_v] = ... get_Cats(new_Cats_v, Cats_nests, bestnest,t,maxiter, w(t),lb,ub); % update the velocities Cats_v(non_parasitized_nests( 1:nCats),:) = new_Cats_v; % ___________________ Evaluation: ___________________________________ % new system newnest = [new_Crows;new_Cuckoos;new_Cats]; [nest,fitness] = get_best_nest(fobj,nest,newnest,fitness); % Find the best objective so far [fitness,I] = sort(fitness); nest = nest(I,:); fmin = fitness(1); bestnest = nest(1,:); Convergence_curve(t) = fmin; % Display the iteration and best optimum obtained so far % if mod(t,10)==0 % display(['At iteration: ', num2str(t), ' Best fitness is: ', num2str(fmin)]); % end end %% End of iterations end function choice = RankingSelection(n, N) ranking = (1:n);% for sorted fitness weights=ranking/sum(ranking); Select_Fitness = cumsum(weights); % (the better nests the higher chance for parasitism) choice=[]; while length(choice) < N Random_Fitness=rand(1,n); selected = ranking(Select_Fitness <= Random_Fitness); choice=union(choice,selected,'stable')'; end end function Sol=simplebounds(Sol,lb,ub) for i=1:size(Sol,1)% return back populations % Sol(i,:)=max(Sol(i,:),lb); % Sol(i,:)=min(Sol(i,:),ub); Vub=Sol(i,:)>ub; Vlb=Sol(i,:)<lb; % Sol(i,:)=(Sol(i,:).*(~(Vub+Vlb)))+ub.*Vub+lb.*Vlb; Sol(i,:)=(Sol(i,:).*(~(Vub+Vlb)))+(lb+rand(size(lb)).*(ub-lb)).*(Vub+Vlb); end end % Growth rates (Section 3.1.2.1, Fig. 3) function [GrowthRateCrows,GrowthRateCats, GrowthRateCuckoos] =Growth_rate(n,maxiter) % According to references [?40-43], and using: % r1=1: The intrinsic growth rate of crows % r2=0.1: The death rate of cuckoos % r3=0.1: The death rate of Cats % d1=0.01:The density-dependent mortality rate of crows % d2=0.01:The density-dependent mortality rate of cuckoos % B1=0.1:The quantity of resources consumed by a cuckoo % B2=0.1:cats? saturation level in predation % c1=0.1:The half-saturation density in predation % c2=0.1:The time lost by cats due to deterrence of cuckoos % Alpha1=0.2:Cuckoos? efficiency in converting their consumption into fitness % Alpha2=0.25:cats? efficiency in converting the predation into fitness % with a crow膨uckoo膨at model: % dx(i)=r1*x(i-1)*(1-c1*x(i-1)-B1*y(i-1)-B2*z(i-1)); % dy(i)=r2*y(i-1)*(-1+Alpha1*x(i-1)-c2*y(i-1)); %#ok<*AGROW> % dz(i)=r3*z(i-1)*(-1+Alpha2*x(i-1)-B3*y(i-1)-c3*z(i-1)); % x(i)=x(i-1)+dx(i); % y(i)=y(i-1)+dy(i); % z(i)=z(i-1)+dz(i); % the growth rates can be obtained as follows: GrowthRateCrows= round(n*linspace(2/3,1/2, maxiter));%Growth rate of Crows GrowthRateCats= round(n*linspace(0.01,1/3, maxiter));%Growth rate of Cats GrowthRateCuckoos=n-GrowthRateCrows-GrowthRateCats;%Growth rate of Cuckoos end % Find the current best nest function [nest,fitness]=get_best_nest(fobj,nest,newnest,fitness) % Evaluating all new solutions for j=1:size(nest,1) fnew = fobj(newnest(j,:)); if fnew<=fitness(j) fitness(j)=fnew; nest(j,:)=newnest(j,:); end end end % 3.1.2.4. Predation phase (crow-cat) % Tracking mode function of Cat Swarm function [Cats, Cats_v] = ... get_Cats(Cats_v, Cats, gx,t,maxiter, w,lb,ub) %velocity limits is modified to be decreased linearly from 1 to 0.25 perCnt = 1-t/maxiter/4; vlb = perCnt*lb; vUp = perCnt*ub; c=2-t/maxiter; % c (social learning) is modified to be decreased linearly from 2 to 1 for iSz2 = 1:size(Cats,1) V_vec = w * Cats_v(iSz2,:) + c*rand*(gx - Cats(iSz2,:)); % Eq. (13) % Apply simple bounds/limits to velocity of cats V_vec = max(V_vec, vlb); V_vec = min(V_vec, vUp); Cats_v(iSz2,:) = V_vec; end Cats = Cats + Cats_v; % Eq. (14) % Apply simple bounds/limits to position of cats Cats = simplebounds( Cats, lb, ub); end % 3.1.2.2. Nesting phase (crow痴 nest) % Levy Crow search algorithm is used to generate new solutions function nest=get_Crows(nest,lb,ub) % Levy flights by Mantegna's algorithm for Crow search algorithm n=size(nest,1); beta=3/2; sigma=0.6965745; % Generation of random candidate crows for following (chasing) num=ceil(n*rand(1,n)); for i=1:n s=nest(i,:); u=randn(size(s))*sigma; v=randn(size(s)); step=u./abs(v).^(1/beta); % Eq.(7) stepsize=0.1*step.*randn(size(s)); % Eq.(8) % Generation of a new position for crow i (state 1) s= s+stepsize.*(nest(num(i),:)-s); % Eq.(6) % Generation of a new position for crow i (state 2)for more exploration out=s<lb | s > ub; s(:,out)=lb(out)+(ub(out)-lb(out)).*rand(1,nnz(out)); % Eq.(9) nest(i,:)=s; end end % 3.1.2.3. Parasitism phase (crow-cuckoo subsystem) % Replace some nests by constructing new solutions/nests function new_nest=get_Cuckoos(nest,lb,ub,t,maxiter) pa=t/2/maxiter; % Adaptive Discovery rate of alien eggs/solutions n=size(nest,1); % probability Eq. (12) K=rand(size(nest)) > pa; % Discovered or not -- a status vector stepsize=rand*(nest(randperm(n),:)-nest(randperm(n),:)); new_nest=nest+stepsize.*K; % Eq. (10) % Apply simple bounds/limits new_nest=simplebounds(new_nest,lb,ub); end function nest =init(n,ub,lb) for i=1:n nest(i,:)=lb+(ub-lb).*rand(size(lb)); % Eq.(5) end end