Increasing sequence similarity to input sequences

[diossynvax]

I've discovered that using random sampling, the highest sequence similarity I typically see to any of the input sequences is around 70-75%. If you wanted to adjust the parameters in order to generate sequences with closer to 80-85% sequence similarity to the input sequences, what would be the best way to alter the parameters in order to do this?