Convergence curves for genetic forward modeling (yellow), and basic Monte Carlo technique (red). To allow equitable comparison between both methods, one generation also corresponds to 50 model evaluations for the Monte Carlo technique. Each curve is an average of 20 runs, and the two sets of thin lines are the slowest and fastest converging solutions (normalized chi^2 less than 2) for both methods. The green line is genetic forward modeling convergence curve obtained using identical parameter settings as for the yellow line, with the single exception that the crossover operator has been turned off within the genetic algorithm