Novel Back Propagation Optimization by Cuckoo Search Algorithm

Algorithm 1

Cuckoo search.

Begin

  Step 1. Initialization.   Set the generation counter 𝐺=1 and the discovery rate 𝑝_𝑎;   initialize

      the population 𝑃  of 𝑛 host nests randomly.

  Step 2.    While   𝐺<𝑀𝑎𝑥𝐺𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛  do

       Sort all the cuckoos.

       Randomly select a cuckoo 𝑖 and implement Lévy flights to replace its solution.

       Evaluate its fitness 𝐹_𝑖.

       Randomly choose another nest 𝑗.

       if  (𝐹_𝑖<𝐹_𝑗)

         Replace j by the new solution.

       end if

       Randomly generate a fraction (𝑝_𝑎) of new nests and replace the worse nests.

       Keep the best nests.

       Sort the population and find the best cuckoo for this generation.

       Pass the current best to the next generation.

       𝐺=𝐺+1.

  Step 3. end while

End.