基于分类的降维多目标进化算法的高维昂贵优化matlab代码.zip

classdef DRMCEAD < ALGORITHM % <multi/many> <real/integer> <expensive> % Dimensionality Reduction-based MCEA/D % delta --- 0.9 --- The probability of choosing parents locally % nr --- 2 --- Maximum number of solutions replaced by each offspring % Rmax --- 10 --- Maximum repeat time of offspring generation % C --- 1.0 --- The parameter for SVM % gamma --- 1.0 --- The parameter for SVM % beta --- 0.5 --- The reduction rate % rho --- 0.1 --- The rate of the superior solutions and the inferior solutions %------------------------------- Reference -------------------------------- % Y. Horaguchi and M. Nakata, High-Dimensional Expensive Optimization by % Classification-based Multiobjective Evolutionary Algorithm with % Dimensionality Reduction, 2023 62nd Annual Conferenc of the Society of % Instrument and Control Engnieers (SICE), 2023, 1535-1542. %------------------------------- Copyright -------------------------------- % Copyright (c) 2023 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". %-------------------------------------------------------------------------- % This function is written by Yuma Horaguchi methods function main(Algorithm, Problem) %% Parameter setting [delta, nr, R_max, C, gamma, beta, rho] = Algorithm.ParameterSet(0.9, 2, 10, 1.0, 1.0, 0.5, 0.1); %% Generate the weight vectors [W, Problem.N] = UniformPoint(Problem.N, Problem.M); %% Detect the neighbours of each solution T = ceil(Problem.N / 10); B = pdist2(W, W); [~, B] = sort(B, 2); B = B(:, 1 : T); %% Initialize population PopDec = UniformPoint(Problem.N, Problem.D, 'Latin'); Population = Problem.Evaluation(repmat(Problem.upper - Problem.lower, Problem.N, 1) .* PopDec + repmat(Problem.lower, Problem.N, 1)); Arc = Population; Z = min(Population.objs, [], 1); sigma = sqrt(1 / (2 * gamma)); %% Optimization while Algorithm.NotTerminated(Arc) % For each sub-problem for i = 1 : Problem.N %% Dimension Reduction IndexDif = DimensionalityReduction(Problem, Arc, beta, rho, W, Z, i); %% Build a SVM classifier C_i = []; label = -1 * ones(length(Arc), 1); for k = 1 : length(B(i, :)) g_A = max(abs(Arc.objs - repmat(Z, length(Arc), 1)) .* W(B(i, k), :), [], 2); [~, rankA] = sort(g_A); for j = 1 : length(Arc) if ~ismember(rankA(j), C_i) C_i = [C_i, rankA(j)]; label(rankA(j)) = +1; break end end end uniform_ADec = (Arc.decs - Problem.lower) ./ (Problem.upper - Problem.lower); svm_mdl = fitcsvm(uniform_ADec(:, IndexDif), label, 'BoxConstraint', C, 'KernelScale', sigma, 'KernelFunction', 'rbf'); %% Choose the parents if rand < delta P = B(i, randperm(end)); else P = randperm(Problem.N); end %% Solution-generation Offspring = DRSolutionGeneration(Problem, Population, IndexDif, P, svm_mdl, R_max, i); %% Evaluate offspring Offspring = Problem.Evaluation(Offspring); %% Update the reference point Z = min(Z, Offspring.obj); %% Update population and archive g_old = max(abs(Population(P).objs - repmat(Z, length(P), 1)) .* W(P, :), [], 2); g_new = max(repmat(abs(Offspring.obj - Z), length(P), 1) .* W(P, :), [], 2); Population(P(find(g_old >= g_new, nr))) = Offspring; Arc = [Arc, Offspring]; %% Check termination criteria Algorithm.NotTerminated(Arc); end end end end end