A New Wind Power Forecasting Approach Based on Conjugated Gradient Neural Network

Algorithm 1

Conjugate gradient descent steps.

(1) Take the initial point

(2)  if  , set as supposed point

   else calculate the Hesse Matrix of

   end if

(3) Calculate the initial conjugated vector by

(4) Find a proper to minimize

(5) Calculate the best step by

   And

(6)  if  

   set as the supposed point

   else if   and

   search a vector in the sub space of

    and , so that

   end if

(7) Set  

(8) Then  

(9) Let  , calculate

(10) if  

   set as the supposed point

   else

   let   and

   then  .

(11) Loop (3).