Application of a simple binary genetic algorithm to a noiseless testbed benchmark

One of the earliest evolutionary computation algorithms, the genetic algorithm, is applied to the noise-free BBOB 2009 testbed. It is adapted to the continuous domain by increasing the number of bits encoding each variable, until a desired resolution is possible to achieve. Good results and scaling are obtained for separable functions, but poor performance is achieved on the other functions, particularly ill-conditioned functions. Overall running times remain fast throughout. Categories and Subject Descriptors

• Type: Conference
• Year: 2009
• Where: GECCO
• Authors: Miguel Nicolau
• Keywords: Earliest Evolutionary Computation, GECCO 2009, Genetic Algorithm, Optimization, Overall Running Times