An Energy Efficient Evolutionary Approach for Smart City-Based IoT Applications

Algorithm 1

Pseudocode for FFNN training using PSO.

Initialization of , , , , ,

While < = do

  1. Map into W and b

  2. Evaluate Equations 11 and 12. This phase is called training of FFNN

  3: Fitness (i.e., FFNN error or MSE) can be obtained using Equation 1.

  if pBestScore > Fitness then

  | pBestScore = Fitness and pBestPosition = x.

  Else

  End

  If gBestScore > fitness then

  | gBestScore = Fitness and gBestPosition = x.

  Else

  End

  4: Now calculate w by using Equation 11.

 5: Update velocity and position of particles according to Equations 11 and 12, respectively.

 End

 Final: PSO’s best particle positions (pBest) are the (W and b) for FFNN.