Optimizing Deep Learning Model for Software Cost Estimation Using Hybrid Meta-Heuristic Algorithmic Approach

Algorithm 2

BAT algorithm (BA).

	For a given objective function: Obj, where i = (1, …, n)
	Initialize the BAT population with the corresponding attributes: velocity , position and pulse frequency
	Initialization features include loudness , maximum number of iterations and acceleration rate
	Repeat step 4 to 15 for every iteration in
	Repeat step 5 to 13 for every single BAT corresponding to
	Evaluate equations (1)–(3) to produce a new solutions
	set
	if (random ) then
	Generate a local solution around one of the chosen best solutions
	end if
	if (random ) (fitness()) then
	Update and
	Increase and reduce (equations (3) and (6))
	end if
	Rank the BATs and find the GlobalBest
	=0