Random walk population sizing as a model of decision making in a deceptive schema partition

Population-sizing models have been shown to be accurate predictors of genetic algorithm convergence. The random walk model incorporates and improves upon previous work in population sizing. This research investigates the ability of the random walk model to predict the underlying genetic mechanisms of schema processing. These mechanisms are evaluated by examining the decisions a genetic algorithm makes in deceptive schema partitions. Results of this research indicate that the random walk model is an accurate predictor of schema partition processing in deceptive partitions. Visualization of the ongoing competition among the schemata in a partition as vertices of a hypercube confirm the dynamics of schema processing predicted by the model. Taken together, the visualization of the process and the statistics of the results help to confirm the random walk model of schema processing at a more detailed level than previous studies of end results only.