Desirability Improvement of Committee Machine to Solve Multiple Response Optimization Problems

Algorithm 1

get Data //include 𝑋, 𝑌 matrixes

setRMSE_network = 1    // beginning of modeling phase

set min_RMSE = 0.4

for all kind of neural networks

   while (RMSE_network > min_RMSE or coefficient of correlation < 0) and iterations < 5

   set 𝑋 and 𝑌

if min_RMSE > RMSE_network

set min_RMSE = RMSE_network

end if

   add one to iterations

   end       // end of while

end for

iteration = 1;

while (iteration < 5)

calculate CM weights using GA randomly

train network

calculate RMSE_network and coefficient of correlation

if RMSE_network < min_RMSE for goal of minimizing in overall RMSE   // end of modeling phase

calculate 𝑋^∗, 𝑦^∗ and GD(𝑋^∗) using GA for goal of maximizing in Global desirability      //

end

   //end