matlab粒子群算法代码实现

粒子群算法（Particle Swarm Optimization，PSO）是一种优化算法，它通过模拟鸟群捕食行为的方式来寻找最优解。下面是用MATLAB实现粒子群算法的代码示例：

function [bestpos, bestfitness] = pso(fobj, dim, lb, ub, maxiter, npop, c1, c2, w)
% fobj: 目标函数
% dim: 变量维度
% lb: 变量下界
% ub: 变量上界
% maxiter: 最大迭代次数
% npop: 粒子数
% c1: 学习因子1
% c2: 学习因子2
% w: 惯性权重

% 初始化粒子群
pos = repmat(lb, npop, 1) + rand(npop, dim) .* repmat((ub-lb), npop, 1);
vel = zeros(npop, dim);
pbestpos = pos;
pbestfitness = ones(npop, 1) * inf;

% 寻找全局最优解
[gbestfitness, gbestid] = min(pbestfitness);
gbestpos = pbestpos(gbestid, :);

% 迭代
for iter = 1:maxiter
    % 更新粒子速度和位置
    vel = w * vel + c1 * rand(npop, dim) .* (pbestpos - pos) + c2 * rand(npop, dim) .* repmat((gbestpos - pos), npop, 1);
    pos = pos + vel;

    % 越界处理
    pos(pos<lb) = lb(pos<lb);
    pos(pos>ub) = ub(pos>ub);

    % 更新每个粒子的最优解
    fitness = feval(fobj, pos');
    update = fitness < pbestfitness;
    pbestpos(update, :) = pos(update, :);
    pbestfitness(update) = fitness(update);

    % 更新全局最优解
    [tempgbestfitness, tempgbestid] = min(pbestfitness);
    if tempgbestfitness < gbestfitness
        gbestfitness = tempgbestfitness;
        gbestpos = pbestpos(tempgbestid, :);
    end
    
    % 输出结果
    fprintf('Iteration %d: Best fitness = %f\n', iter, gbestfitness);
end

bestpos = gbestpos;
bestfitness = gbestfitness;
end

使用方法：

定义一个目标函数，例如：

function y = sphere(x)
y = sum(x.^2);
end

然后调用`pso`函数：

fobj = @sphere;
dim = 10;
lb = -100;
ub = 100;
maxiter = 100;
npop = 50;
c1 = 2;
c2 = 2;
w = 0.7;
[bestpos, bestfitness] = pso(fobj, dim, lb, ub, maxiter, npop, c1, c2, w);

其中，`dim`是变量维度，`lb`和`ub`分别是变量的下界和上界，`maxiter`是最大迭代次数，`npop`是粒子数，`c1`和`c2`是学习因子，`w`是惯性权重。`bestpos`是最优解，`bestfitness`是最优解对应的目标函数值。