The E ects of Randomly Sampled Training Data on

The eeects of randomly sampled training data on genetic programming performance is empirically investigated. Often the most natural, if not only, means of characterizing the target behaviour for a problem is to randomly sample training cases inherent to that problem. A natural question to raise about this strategy is, how deleterious is the randomly sampling of training data to evolution performance? Will sampling reduce the evolutionary search to hill climbing? Can re-sampling during the run be advantageous? We address these questions by undertaking a suite of diierent GP experiments. Parameters include various sampling strategies (single, re-sampling, ideal samples), genera-tional and steady{state evolution, and non{ evolutionary strategies such as hill climbing and random search. The experiments connrm that random sampling eeectively characterizes stochastic domains during genetic programming , provided that a suuciently representative sample is used. An unexpected result is that genetic programming may perform worse than random search when the sampled training sets are exceptionally poor. We conjecture that poor training sets cause evolution to prematurely converge to undesirable optima, which irrevocably handicaps the population's diversity and viability.