Reduce computational time for nested for loops

Hi all

I wondered how could I reduce computational time for this for loop:

for i=1:length(Lungh)
    for j=i:length(Largh)
        for k=1:length(T)
            for w=1:length(Delta) 
                    for av=1:length(A_deg_prua_verticale)
                        for h=1:length(H)
                            for b=1:length(B)
                                for ti=1:length(t)
    if abs(CB(i,j,k,w)>0)
        if Largh(j)<Lungh(i)
               if T(k)<Lungh(i)
    L_cb(i)=Lungh(i);
    Larg_cb(j)=Largh(j);
    T_cb(k)=T(k);
    Dead_rise(k,j)=atand(Largh(j)/(2* T_cb(k)));
    Delta_cb(w)=Delta(w);
    
    H_triangolo(k,av)=(T_cb(k))*tand(A_deg_prua_verticale(av));
     
    Aw_d(i,k,av)=(L_cb(i)-H_triangolo(k,av))*Larg_cb(j)+(Larg_cb(j)* H_triangolo(k,av)/2);
    Awc(i,k,av)=2*Aw_d(i,k,av);
    Cw(i,j,k,av)= Aw_d(i,k,av)/(L_cb(i)*Larg_cb(j));
    Zc(i,j,k,w,av)=T_cb(k)/(1+(CB(i,j,k,w)/Cw(i,j,k,av)));
    Ix_demihull(i,j,k,w,av)=(W_tonn(w)/(2*gamma))*((Cw(i,j,k,av)^2)*...
        (Larg_cb(j)^2))/(11.4*CB(i,j,k,w)*T_cb(k));
    Ix_upper(j,h,ti,b)=(B(b)-2*Larg_cb(j))*(((t(ti)^3)/3)+t(ti)*H(h)^2-H(h)*t(ti)^2);
    Ix(i,j,k,w,av,h,ti,b)=2*Ix_demihull(i,j,k,w,av)+Ix_upper(j,h,ti,b);
    
    ZG(j,k,h,ti,b)=((Largh(j)*H(h)^2)-(Largh(j)*T_cb(k)^2)+(B(b)*(H(h)^2)/2)-B(b)*(H(h)^2+t(ti)^2-2*H(h)*t(ti))/2-(Largh(j)*H(h)^2)+...
        Largh(j)*(H(h)^2+t(ti)^2-2*H(h)*t(ti))+4*tand(Dead_rise(k,j))*(T_cb(k)^3)/3)/...
        ((Largh(j)*T_cb(k)^2)+2*Largh(j)*(H(h)-t(ti)-T_cb(k))+(2*((B(b)/2)-Largh(j))*t(ti)));
     
    Iy(i,j,k,w,av)=0.07*(Cw(i,j,k,av)/9.81)*W_tonn(w)*L_cb(i)^2;
    Delta_Iy(i,j,k,av)=0.055*(gamma/9.81)*(Larg_cb(j)^2)*...
        (L_cb(i)^3)*(Cw(i,j,k,av)^2)/((3-2*Cw(i,j,k,av))*(3-Cw(i,j,k,av)));
    if  Zc(i,j,k,w,av)<ZG(j,k,h,ti,b)&&Zc(i,j,k,w,av)>0&&ZG(j,k,h,ti,b)>0&&ZG(j,k,h,ti,b)<H(h)&&Zc(i,j,k,w,av)<H(h)
        
        Z_C(i,j,k,w,av)=Zc(i,j,k,w,av);
        Z_G(j,k,h,ti,b)=ZG(j,k,h,ti,b);
        
    end
    R(i,j,k,w,av)=(Cw(i,j,k,av)^2)*(L_cb(i)^2)/(14*CB(i,j,k,w)*T_cb(k));
    H_L(i,j,k,w,av,h,ti,b)=R(i,j,k,w,av)+Z_C(i,j,k,w,av)-Z_G(j,k,h,ti,b);
    
    kd(j,b)=((B(b)/2)-Largh(j))+(Largh(j)/2);
    GM(i,j,k,w,av,h,ti,b)=(2*gamma*(Ix_demihull(i,j,k,w,av)+(kd(j,b)^2)*...
        (Aw_d(i,k,av)))/(W_tonn(w)))+Z_C(i,j,k,w,av)-Z_G(j,k,h,ti,b);
    
    if GM(i,j,k,w,av,h,ti,b)>0&&H_L(i,j,k,w,av,h,ti,b)>0
        
    T_teta(i,j,k,w,av,h,ti,b)=(1/(2*pi))*(sqrt(Ix(i,j,k,w,av,h,ti,b)/(W_tonn(w)*GM(i,j,k,w,av,h,ti,b)))) ;
    T_p(i,j,k,w,av,h,ti,b)=2*pi*sqrt((Iy(i,j,k,w,av))/(W_tonn(w)*H_L(i,j,k,w,av,h,ti,b)));
    T_heaving(i,k,w,av)=2*pi*sqrt(W_tonn(w)/(9.81*gamma*Awc(i,k,av)));
    
    if T_teta(i,j,k,w,av,h,ti,b)>0.2
        T_teta_finale(i,j,k,w,av,h,ti,b)=T_teta(i,j,k,w,av,h,ti,b); 
        T_p_finale(i,j,k,w,av,h,ti,b)=T_p(i,j,k,w,av,h,ti,b);
        T_heaving_finale(i,k,w,av)=T_heaving(i,k,w,av);
    end
    end
        
               end
        end
    end
                                end
                            end
                        end
                    end
            end
        end
    end
end

where length(Lungh) = 6, length(Largh)=9, length(T)=5, length(Delta)=7, length(A_deg_prua_verticale)=9, length(H)=11, length(B)=3, length(t)=3

Thank you for the help

Regards!

Answer 1

[Sharmin Kibria]

Hi,

You can definitely speed up the process by vectorization.

https://www.mathworks.com/help/matlab/matlab_prog/vectorization.html

Replacing the loops with vectorized solution can speed up the computation.

Thanks

Sharmin

[EldaEbrithil]

Thank you very much Sharmin!