1 简介
针对目前农村公路造价估算存在的问题,结合神经网络和麻雀搜索算法各自的特点,提出了一种基于改进型麻雀搜索算法优化BP神经网络(improved sparrow search agorithm optimized BP neural network,Tent-SSA-BP)的模型,对公路造价进行预测.为提高预测算法的搜索和跳出局部最优的能力,在标准麻雀搜索算法(sparrow search algorithm,SSA)中借助混沌的思想,使加入者以一定的概率向发现者靠近,缩短了算法的运行时间,保证了全局收敛和种群的多样性.最后将算法模型与BP神经网络模型和麻雀搜索算法优化BP神经网络(sparrow search algorithm BP,SSA-BP)模型进行对比,使用三种评价指标对三种算法模型进行评价,结果表明的tent-SSA-BP模型能够及时准确的公路数据进行预测,ISSA-BP模型的均方误差(MSE),平均绝对误差(MAE)和决定系数(R2)分别为2.511E-11,1.969E-06和0.99989.
2 部分代码
function [fMin , bestX, Convergence_curve] = SSA(X, N, M, c, d, dim, fobj)
P_percent = 0.2; % 发现者的种群规模占总种群规模的百分比
pNum = round(N*P_percent); % 发现者数量20%
SD = pNum/2; % 警戒者数量10%
ST = 0.8; % 安全阈值
lb = c.*ones(1, dim); % 下限
ub = d.*ones(1,dim); % 上限
% 初始化
for i = 1:N
% X(i, :) = lb + (ub - lb) .* rand(1, dim);
fitness(i) = fobj(X(i, :));
end
pFit = fitness;
pX = X; % 与pFit相对应的个体最佳位置
[fMin, bestI] = min(fitness); % fMin表示全局最优解
bestX = X(bestI, :); % bestX表示全局最优位置
%% 迭代寻优
for t = 1 : M
[~, sortIndex] = sort(pFit); % 排序
[fmax, B] = max(pFit);
worst = X(B, :);
%% 发现者位置更新
r2 = rand(1);
if r2 < ST
for i = 1:pNum % Equation (3)
r1 = rand(1);
X(sortIndex(i), :) = pX(sortIndex(i), :)*exp(-(i)/(r1*M));
X(sortIndex(i), :) = Bounds(X(sortIndex(i), :), lb, ub);
fitness(sortIndex(i)) = fobj(X(sortIndex(i), :));
end
else
for i = 1:pNum
X(sortIndex(i), :) = pX(sortIndex(i), :)+randn(1)*ones(1, dim);
X(sortIndex(i), :) = Bounds(X(sortIndex(i), :), lb, ub);
fitness(sortIndex(i)) = fobj(X(sortIndex(i), :));
end
end
[~, bestII] = min(fitness);
bestXX = X(bestII, :);
%% 跟随者位置更新
for i = (pNum+1):N % Equation (4)
A = floor(rand(1, dim)*2)*2-1;
if i > N/2
X(sortIndex(i), :) = randn(1)*exp((worst-pX(sortIndex(i), :))/(i)^2);
else
X(sortIndex(i), :) = bestXX+(abs((pX(sortIndex(i), :)-bestXX)))*(A'*(A*A')^(-1))*ones(1, dim);
end
X(sortIndex(i), :) = Bounds(X(sortIndex(i), :), lb, ub);
fitness(sortIndex(i)) = fobj(X(sortIndex(i), :));
end
%% 警戒者位置更新
c = randperm(numel(sortIndex));
b = sortIndex(c(1:SD));
for j = 1:length(b) % Equation (5)
if pFit(sortIndex(b(j))) > fMin
X(sortIndex(b(j)), :) = bestX+(randn(1, dim)).*(abs((pX(sortIndex(b(j)), :) -bestX)));
else
X(sortIndex(b(j)), :) = pX(sortIndex(b(j)), :)+(2*rand(1)-1)*(abs(pX(sortIndex(b(j)), :)-worst))/(pFit(sortIndex(b(j)))-fmax+1e-50);
end
X(sortIndex(b(j)), :) = Bounds(X(sortIndex(b(j)), :), lb, ub);
fitness(sortIndex(b(j))) = fobj(X(sortIndex(b(j)), :));
end
for i = 1:N
% 更新个体最优
if fitness(i) < pFit(i)
pFit(i) = fitness(i);
pX(i, :) = X(i, :);
end
% 更新全局最优
if pFit(i) < fMin
fMin = pFit(i);
bestX = pX(i, :);
end
end
Convergence_curve(t) = fMin;
disp(['SSA: At iteration ', num2str(t), ' ,the best fitness is ', num2str(fMin)]);
end
%% 边界处理
function s = Bounds(s, Lb, Ub)
% 下界
temp = s;
I = temp < Lb;
temp(I) = Lb(I);
% 上界
J = temp > Ub;
temp(J) = Ub(J);
% 更新
s = temp;
3 仿真结果
4 参考文献
[1]许亮, 张紫叶, 陈曦,等. 基于改进麻雀搜索算法优化BP神经网络的气动光学成像偏移预测[J]. 光电子.激光, 2021, 32(6):6.
博主简介:擅长智能优化算法、神经网络预测、信号处理、元胞自动机、图像处理、路径规划、无人机等多种领域的Matlab仿真,相关matlab代码问题可私信交流。
部分理论引用网络文献,若有侵权联系博主删除。