Using a threshold of 50 kips, we ran 20 trials with a limit of 500 validity evaluations, and 18 found satisfactory solutions, of which 15 were further refined into valid solutions. The weights of the solutions ranged from 13.45 to 25.15, with a mean of 18.41 and a median of 19.11. Limited experiments suggested that a still lower threshold produced even better quality (lower weight)solutions. A designer might not know of a tighter bound on the weight initially, but we would expect a tool like this to be applied iteratively.