For gn (gene) = 1 : 30 |
Initialize the population para with host nests (solutions) with |
dimension = 3 that is . |
For : 10 |
Calculate fob (fitness value) for all 10 solutions using RNN and FPA; |
If () |
Break; |
End if; |
End for; |
If (minimum (fob) ) |
For : max iteration (100) |
Randomly select a cuckoo (th) avoid current best; |
Randomly generate another nest (th) keeping current best nests by Lévy flights; |
If () |
is replaced by the news solution; |
End if; |
If () |
Break; |
End; |
Discard the worse nests with a fractional probability (); |
Keep the highest quality nest that is best solution with best fitness value; |
Rank the available solutions and locate the current best; |
End for; |
End if; |
End for; |
Post-processing and visualization of GRN; |
Function fob (para, gn) |
Initialize a rnnpara population of nf (30) with dimension df (5) pollen randomly with a switch probability |
Find the fitness fun for all solutions and best pollen among them |
For (tf < MaxGeneration (2000)) |
For : nf |
If , |
Draw (df-dimensional) step vector L which obeys a Lévy distribution |
Global pollination via equation (4) |
Else |
Draw from a uniform distribution in |
Randomly choose jf and kf among all the solutions |
Do local pollination via equation (5) |
End if |
Evaluate fitness fun of new solutions of pollens |
If new solutions are better |
Update them in the population |
End if |
End for |
Find store the best fitness of current iteration |
If ((bestfitness (tf) <) (() && ((bestfitness () − bestfitness (tf)) < ))) |
Break |
End if |
End for |
Return bestfitness |
End fob |
Function fun (rnnpara, para, gn) |
Define (5) times series data with (50) sample point |
Calculate the gene expression value of next time instance using equation (1) |
Determine the squared error using equation (7) |
Return error |
End fun |