【LSTM回归预测】基于北方苍鹰算法优化长短时记忆网络NGO-LSTM实现风电数据预测附matlab代码...

✅作者简介：热爱科研的Matlab仿真开发者，修心和技术同步精进，matlab项目合作可私信。

🍎个人主页：Matlab科研工作室

🍊个人信条：格物致知。

⛄ 内容介绍

针对海上风电功率预测精度差的问题,提出一种NGO-LSTM模型.选择在时间序列问题处理上具有良好性能的长短期记忆(LSTM)神经网络,通过寻优能力强、收敛速度快的北方苍鹰算法对LSTM网络隐含层神经元个数、学习率和训练次数等超参数进行优化,得到NGO-LSTM模型.采用江苏省盐城市某400 MW风电场功率数据进行算例分析,在不同条件变量下分别使用NGO-LSTM模型、LSTM模型预测,仿真结果表明,NGO-LSTM模型具有更高的预测精度、更好的预测稳定性.

⛄ 部分代码

%%

% NGO.

% Northern Goshawk Optimization: A New Swarm-Based Algorithm for Solving Optimization Problems

% Mohammad Dehghani1, Pavel Trojovský1, and Stepan Hubálovský2

% 1Department of Mathematics, Faculty of Science, University of Hradec Králové, 50003 Hradec Králové, Czech Republic

% 2Department of Applied Cybernetics, Faculty of Science, University of Hradec Králové, 50003 Hradec Králové, Czech Republic

% " Optimizer"

%%

function [Score,Best_pos,NGO_curve]=NGO(Search_Agents,Max_iterations,Lowerbound,Upperbound,dimensions,objective)

tic

disp('PLEASE WAIT, The program is running.')

Lowerbound=ones(1,dimensions).*(Lowerbound);                              % Lower limit for variables

Upperbound=ones(1,dimensions).*(Upperbound);                              % Upper limit for variables

X=[];

X_new=[];

fit=[];

fit_new=[];

NGO_curve=zeros(1,Max_iterations);

%%

for i=1:dimensions

    X(:,i) = Lowerbound(i)+rand(Search_Agents,1).*(Upperbound(i) -Lowerbound(i));              % Initial population

end

for i =1:Search_Agents

    L=X(i,:);

    fit(i)=objective(L);                    % Fitness evaluation (Explained at the top of the page. )

end

for t=1:Max_iterations  % algorithm iteration

    %%  update: BEST proposed solution

    [best , blocation]=min(fit);

    if t==1

        xbest=X(blocation,:);                                           % Optimal location

        fbest=best;                                           % The optimization objective function

    elseif best<fbest

        fbest=best;

        xbest=X(blocation,:);

    end

    %% UPDATE Northern goshawks based on PHASE1 and PHASE2

    for i=1:Search_Agents

        %% Phase 1: Exploration

        I=round(1+rand);

        k=randperm(Search_Agents,1);

        P=X(k,:); % Eq. (3)

        F_P=fit(k);

        if fit(i)> F_P

            X_new(i,:)=X(i,:)+rand(1,dimensions) .* (P-I.*X(i,:)); % Eq. (4)

        else

            X_new(i,:)=X(i,:)+rand(1,dimensions) .* (X(i,:)-P); % Eq. (4)

        end

        X_new(i,:) = max(X_new(i,:),Lowerbound);X_new(i,:) = min(X_new(i,:),Upperbound);

        % update position based on Eq (5)

        L=X_new(i,:);

        fit_new(i)=objective(L);

        if(fit_new(i)<fit(i))

            X(i,:) = X_new(i,:);

            fit(i) = fit_new(i);

        end

        %% END PHASE 1

        %% PHASE 2 Exploitation

        R=0.02*(1-t/Max_iterations);% Eq.(6)

        X_new(i,:)= X(i,:)+ (-R+2*R*rand(1,dimensions)).*X(i,:);% Eq.(7)

        X_new(i,:) = max(X_new(i,:),Lowerbound);X_new(i,:) = min(X_new(i,:),Upperbound);

        % update position based on Eq (8)

        L=X_new(i,:);

        fit_new(i)=objective(L);

        if(fit_new(i)<fit(i))

            X(i,:) = X_new(i,:);

            fit(i) = fit_new(i);

        end

        %% END PHASE 2

    end% end for i=1:N

    %%

    %% SAVE BEST SCORE

    best_so_far(t)=fbest; % save best solution so far

    average(t) = mean (fit);

    Score=fbest;

    Best_pos=xbest;

    NGO_curve(t)=Score;

end

%%

d

⛄ 运行结果

⛄ 参考文献

[1]李森文, 张伟, 李纯宇,等. 基于SSA-LSTM的海上风电功率预测[J]. 机械与电子, 2022(040-006).

[1]杨耘, 王彬泽, 刘艳,等. 基于时空优化LSTM深度学习网络的气温预测[J]. 徐州工程学院学报：自然科学版, 2020, 35(2):6.

⛳️ 完整代码

❤️部分理论引用网络文献，若有侵权联系博主删除

❤️ 关注我领取海量matlab电子书和数学建模资料