用matlab遗传算法pso优化支持向量机分类代码

以下是一个用Matlab编写的支持向量机分类器的遗传算法和粒子群优化算法的优化代码：

遗传算法：

% 遗传算法参数初始化
popsize = 30; % 种群大小
maxiter = 100; % 最大迭代次数
pc = 0.8; % 交叉概率
pm = 0.01; % 变异概率
N = 10; % 变量个数
Lb = zeros(1,N); % 变量下限
Ub = ones(1,N); % 变量上限

% 初始化种群
pop = Lb + (Ub - Lb) .* rand(popsize,N);

% 迭代
for iter = 1:maxiter
    % 计算适应度
    fitness = zeros(popsize,1);
    for i = 1:popsize
        fitness(i) = svm_classify(pop(i,:));
    end
    
    % 选择
    [fitness,idx] = sort(fitness,'descend');
    pop = pop(idx,:);
    pop_new = zeros(popsize,N);
    for i = 1:popsize
        j = mod(i-1,popsize)+1;
        k = mod(i,popsize)+1;
        if rand < pc % 交叉
            alpha = rand(1,N);
            pop_new(i,:) = alpha .* pop(j,:) + (1-alpha) .* pop(k,:);
        else % 不交叉
            pop_new(i,:) = pop(j,:);
        end
        if rand < pm % 变异
            pop_new(i,:) = pop_new(i,:) + (Ub-Lb) .* rand(1,N);
        end
    end
    pop = pop_new;

end

% 找出最佳解
fitness = zeros(popsize,1);
for i = 1:popsize
    fitness(i) = svm_classify(pop(i,:));
end
[fitness,idx] = sort(fitness,'descend');
best = pop(idx(1),:);

粒子群优化算法：

% PSO参数初始化
popsize = 30; % 粒子群大小
maxiter = 100; % 最大迭代次数
w = 0.8; % 惯性权重
c1 = 2; % 个体认知因子
c2 = 2; % 群体社会因子
N = 10; % 变量个数
Lb = zeros(1,N); % 变量下限
Ub = ones(1,N); % 变量上限

% 初始化粒子群
pop = Lb + (Ub - Lb) .* rand(popsize,N);
vel = rand(popsize,N);

% 迭代
for iter = 1:maxiter
    % 计算适应度
    fitness = zeros(popsize,1);
    for i = 1:popsize
        fitness(i) = svm_classify(pop(i,:));
    end
    
    % 找出最佳解
    [best_val,idx] = max(fitness);
    best = pop(idx,:);
    
    % 更新速度和位置
    for i = 1:popsize
        vel(i,:) = w * vel(i,:) + c1 * rand(1,N) .* (pop(i,:) - pop_best) + c2 * rand(1,N) .* (best - pop(i,:));
        pop(i,:) = pop(i,:) + vel(i,:);
    end
    
    % 限制位置在边界内
    pop(pop < Lb) = Lb(pop < Lb);
    pop(pop > Ub) = Ub(pop > Ub);
end

% 找出最佳解
fitness = zeros(popsize,1);
for i = 1:popsize
    fitness(i) = svm_classify(pop(i,:));
end
[fitness,idx] = sort(fitness,'descend');
best = pop(idx(1),:);

其中， `svm_classify` 是支持向量机分类器的函数，需要自己定义。