Implement classical evolutionary algorithm – DIFFERENTIAL EVOLUTION (DE) AND further its simple parameter auto-propagation strategy jDE.
Algorithms to implement:
• DE/Rand/1/Bin (classical evolutionary algorithm) and
• jDE
Test functions:
• 1st DeJong function
• 2nd DeJong function
• Schweffel function
• Rastrigin
• Ackley I/II
Both algorithms must be repeated 30 times for each test function – to obtain some statistical
background – you will calculate (from 30 final best results) Min, Max, Mean, Median, and Std.
Dev. Values for each test function. You also have to confirm your findings by plotting of best
solution from each iteration – i.e. convergence graph. Your task is to plot:
• Convergence plot of all 30 runs in one graph (30 lines in 1 graph) – totally 10 plots (2
algorithms x 5 functions)
• Convergence plot of average best sol – i.e. average best solution in each iteration
(one line) – totally 10 plots (2 algorithms x 5 functions)
This is very similar to my computer modeling evolution and ecology classes I have taken at UCSD, which I received an A for. I have the biology background for these models and I have the programming skills for the code.