[dogma1138]

Why would we need to be able to specifically introduce favorable mutations rather than just induce random mutation and select for favorable ones?

We’ve been taking advantage of the latter with selective breeding for millennia and more recently through inducing random mutations with techniques such as irradiation the green revolution was quite famously kick started by new cultivars that where developed by irradiating seeds and selecting those which led to plants with favorable characteristics.

[stultissimus]

In general we can conduct 'directed evolution' (selective breeding is an example). In this example, however, we're talking about improving the resistance of Halderia to the possible future mutations the viruses could develop to exploit them. This is a much, much larger harder to search space; the cost function is not well defined.

[eru]

Well, whenever your population of halderia is big enough, you can dial up the radiation a bit. Most mutations will be bad, some might be good.

When you need the population size to recover, you dial down the radiation a bit.