The well-known Griewangk function, used for evaluation of evolutionary algorithms, becomes easier as the number of dimensions grows. This paper suggests three alternative implementations that maintain function complexity for highdimensional versions of the problem. Diagonal slices of the search landscape and local search are used to demonstrate and evaluate the difficulty of each function. Categories and Subject Descriptors I.2.8 [Artificial Intelligence]: Problem Solving, Control
Artem Sokolov, L. Darrell Whitley, Monte Lunacek