【BP预测】基于人工蜂群算法优化BP神经网络实现数据预测附matlab代码

1 简介

针对BP网络水质评价模型的不足,引入人工蜂群(ABC)算法,将求解BP神经网络各层权值、阀值的过程转化为蜜蜂寻找最佳蜜源的过程,提出了一种新的结合人工蜂群算法的BP网络水质评价方法(ABC-BP)。并以2000—2006年渭河监测断面的10组实测数据作为测试样本对其水质进行了评价,实验结果表明该方法得到的水质评价结果准确,并具有很强的稳定性和鲁棒性。

2 部分代码

function [bestsol,yy] = ABC(prob,lb,ub,Np,T,limit)
%% Starting of ABC
f = NaN(Np,1); % Vector to store the objective function value of the population members
fit = NaN(Np,1); % Vector to store the fitness function value of the population members
trial = NaN(Np,1); % Initialization of the trial vector
D = length(lb); % Determining the number of decision variables in the problem
P = repmat(lb,Np,1) + repmat((ub-lb),Np,1).*rand(Np,D); % Generation of the initial population
for p = 1:Np
f(p) = prob(P(p,:)); % Evaluating the objective function value
fit(p) = CalFit(f(p)); % Evaluating the fitness function value
end
[bestobj, ind] = min(f); % Determine and memorize the best objective value
bestsol = P(ind,:); % Determine and memorize the best solution
bestobj1=1000;
for t = 1:T
%% Employed Bee Phase
for i = 1:Np
[trial,P,fit,f] = GenNewSol(prob, lb, ub, Np, i, P, fit, trial, f, D);
end
%% Onlooker Bee Phase
% as per the code of the inventors available at https://abc.erciyes.edu.tr/
% prob=(0.9.*Fitness./max(Fitness))+0.1;
% MATLAB Code of the ABC algorithm version 2 has been released (14.12.2009) (more optimized coding)
probability = 0.9 * (fit/max(fit)) + 0.1;
m = 0; n = 1;
while(m < Np)
if(rand < probability(n))
[trial,P,fit,f] = GenNewSol(prob, lb, ub, Np, n, P, fit, trial, f, D);
m = m + 1;
end
n = mod(n,Np) + 1;
end
[bestobj,ind] = min([f;bestobj]);
CombinedSol = [P;bestsol];
bestsol = CombinedSol(ind,:);
if bestobj<bestobj1
yy(t)=bestobj;
else
yy(t)=bestobj1;
end
%% Scout Bee Phase
[val,ind] = max(trial);
if (val > limit)
trial(ind) = 0; % Reset the trial value to zero
P(ind,:) = lb + (ub-lb).*rand(1,D); % Generate a random solution
f(ind) = prob(P(ind,:)); % Determine the objective function value of the newly generated solution
fit(ind) = CalFit(f(ind)); % Determine the fitness function value of the newly generated solution
end
end
[bestfitness,ind] = min([f;bestobj]);
CombinedSol = [P;bestsol];
bestsol = CombinedSol(ind,:);

3 仿真结果

​4 参考文献

[1]苏彩红, 向娜, 陈广义,等. 基于人工蜂群算法与BP神经网络的水质评价模型[J]. 环境工程学报, 2012.

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。