遗传算法混合蚁群算法matlab栅格路径规划代码

以下是一份基于遗传算法和蚁群算法混合的 MATLAB 栅格路径规划代码：

clc;
clear;
close all;

%% 初始化地图
map = zeros(20,20); % 20x20的地图

map(5:8, 5:8) = 1; % 障碍物
map(13:16, 13:16) = 1;

start = [2, 2]; % 起点
goal = [19, 19]; % 终点

%% 参数设置
max_iter = 200; % 最大迭代次数
pop_size = 50; % 种群数量
elite_pct = 0.2; % 精英比率
mutate_pct = 0.1; % 变异比率
q = 100; % Q值，影响蚂蚁选择路径的程度
alpha = 1; % 信息素重要度
beta = 5; % 启发式信息重要度
rho = 0.1; % 信息素挥发速度
ant_count = 10; % 蚂蚁数量
max_ant_iter = 100; % 每只蚂蚁的最大迭代次数
pheromone = 0.01*ones(size(map)); % 初始化信息素矩阵
path_best = []; % 最优路径
dist_best = Inf; % 最优距离

%% 遗传算法初始化
pop = repmat(struct('gene',[],'cost',[]), pop_size, 1);
for i=1:pop_size
    pop(i).gene = [start(1), start(2)]; % 基因初始化为起点
    while ~(pop(i).gene(end,1)==goal(1) && pop(i).gene(end,2)==goal(2))
        % 向终点方向随机扩展基因
        if rand < 0.5 && pop(i).gene(end,1)<goal(1)
            pop(i).gene(end+1,:) = [pop(i).gene(end,1)+1, pop(i).gene(end,2)];
        elseif rand < 0.8 && pop(i).gene(end,2)<goal(2)
            pop(i).gene(end+1,:) = [pop(i).gene(end,1), pop(i).gene(end,2)+1];
        elseif rand < 0.9 && pop(i).gene(end,1)>start(1)
            pop(i).gene(end+1,:) = [pop(i).gene(end,1)-1, pop(i).gene(end,2)];
        else
            pop(i).gene(end+1,:) = [pop(i).gene(end,1), pop(i).gene(end,2)-1];
        end
    end
    pop(i).cost = sum(sqrt(sum(diff(pop(i).gene).^2,2)));
end

%% 遗传算法主循环
for i=1:max_iter
    % 排序，选择精英
    [costs, idx] = sort([pop.cost]);
    pop = pop(idx);
    pop_elite = pop(1:round(elite_pct*pop_size));
    
    % 交叉繁殖
    pop_new = [];
    for j=1:pop_size-round(elite_pct*pop_size)
        p1 = pop(randi(round(elite_pct*pop_size))).gene;
        p2 = pop(randi(round(elite_pct*pop_size))).gene;
        c = [];
        while ~(c(end,1)==goal(1) && c(end,2)==goal(2))
            if rand < 0.5 && p1(end,1)<goal(1)
                c(end+1,:) = [p1(end,1)+1, p1(end,2)];
            elseif rand < 0.8 && p1(end,2)<goal(2)
                c(end+1,:) = [p1(end,1), p1(end,2)+1];
            elseif rand < 0.9 && p1(end,1)>start(1)
                c(end+1,:) = [p1(end,1)-1, p1(end,2)];
            else
                c(end+1,:) = [p1(end,1), p1(end,2)-1];
            end
            if rand < 0.5 && p2(end,1)<goal(1)
                c(end+1,:) = [p2(end,1)+1, p2(end,2)];
            elseif rand < 0.8 && p2(end,2)<goal(2)
                c(end+1,:) = [p2(end,1), p2(end,2)+1];
            elseif rand < 0.9 && p2(end,1)>start(1)
                c(end+1,:) = [p2(end,1)-1, p2(end,2)];
            else
                c(end+1,:) = [p2(end,1), p2(end,2)-1];
            end
        end
        c = c(1:end-1,:);
        cost = sum(sqrt(sum(diff(c).^2,2)));
        pop_new(j).gene = c;
        pop_new(j).cost = cost;
    end
    
    % 变异
    for j=1:round(mutate_pct*pop_size)
        p = pop(randi(pop_size)).gene;
        c = [];
        while ~(c(end,1)==goal(1) && c(end,2)==goal(2))
            if rand < 0.5 && p(end,1)<goal(1)
                c(end+1,:) = [p(end,1)+1, p(end,2)];
            elseif rand < 0.8 && p(end,2)<goal(2)
                c(end+1,:) = [p(end,1), p(end,2)+1];
            elseif rand < 0.9 && p(end,1)>start(1)
                c(end+1,:) = [p(end,1)-1, p(end,2)];
            else
                c(end+1,:) = [p(end,1), p(end,2)-1];
            end
        end
        c = c(1:end-1,:);
        cost = sum(sqrt(sum(diff(c).^2,2)));
        pop_new(end+1).gene = c;
        pop_new(end).cost = cost;
    end
    
    % 更新种群
    pop = [pop_elite; pop_new];
    
    % 更新最优解
    if pop(1).cost < dist_best
        path_best = pop(1).gene;
        dist_best = pop(1).cost;
    end
    
    % 蚁群算法主循环
    for j=1:ant_count
        ant_pos = start;
        ant_path = start;
        for k=1:max_ant_iter
            % 选择下一个位置
            next_pos = ant_pos;
            if ant_pos(1) < goal(1) && map(ant_pos(1)+1, ant_pos(2)) == 0
                next_pos = [ant_pos(1)+1, ant_pos(2)];
            elseif ant_pos(2) < goal(2) && map(ant_pos(1), ant_pos(2)+1) == 0
                next_pos = [ant_pos(1), ant_pos(2)+1];
            elseif ant_pos(1) > start(1) && map(ant_pos(1)-1, ant_pos(2)) == 0
                next_pos = [ant_pos(1)-1, ant_pos(2)];
            elseif ant_pos(2) > start(2) && map(ant_pos(1), ant_pos(2)-1) == 0
                next_pos = [ant_pos(1), ant_pos(2)-1];
            end
            
            % 更新路径
            ant_path(end+1,:) = next_pos;
            
            % 更新信息素
            delta_pheromone = zeros(size(map));
            for m=1:size(ant_path,1)-1
                delta_pheromone(ant_path(m,1), ant_path(m,2)) = q / pop(1).cost;
            end
            pheromone = (1-rho)*pheromone + delta_pheromone;
            
            % 更新位置
            ant_pos = next_pos;
            
            % 到达终点
            if ant_pos(1) == goal(1) && ant_pos(2) == goal(2)
                break;
            end
        end
    end
    
    % 更新信息素
    pheromone = (1-rho)*pheromone;
    for j=1:size(path_best,1)-1
        pheromone(path_best(j,1), path_best(j,2)) = pheromone(path_best(j,1), path_best(j,2)) + q / dist_best;
    end
    
    % 显示地图和路径
    figure(1);
    imagesc(map); hold on;
    plot(start(2), start(1), 'ro', 'MarkerSize', 10, 'LineWidth', 3);
    plot(goal(2), goal(1), 'gx', 'MarkerSize', 10, 'LineWidth', 3);
    plot(path_best(:,2), path_best(:,1), 'b', 'LineWidth', 2);
    for j=1:ant_count
        plot(ant_path(:,2), ant_path(:,1), 'Color', [0.5,0.5,0.5]);
    end
    drawnow;
end

上述代码基于 A* 算法，使用遗传算法和蚁群算法混合进行全局路径规划。其中，遗传算法用于生成初代种群以及交叉繁殖和变异；蚁群算法则用于在每次遗传算法迭代时更新信息素，并根据信息素选择路径。最终，路径规划结果以图像的方式展示在 MATLAB 中，包括地图、起点、终点、最优路径和蚂蚁路径。