Our drat picks an end she wishes to move towards her house. She then goes to the other end and puts a block on that end to keep it from moving backwards. She then goes back to the end she wishes to move and she can just push the log. The log will rotate around the block she placed against the far end.
When she has pushed the one end far enough, she reverses herself and begins to push the formerly fixed end.
It turns out the the log acts as a lever. Since the drat travels exactly twice as far as the center of the log, the force she needs to overcome friction is k * 1/2mg or 1/10 mg, which is exactly what she can lift. (And where does the other half of the force come from? The block she placed on the far end of the log).
*******Addendum
Our Drat is an extraordinarily smart drat. Not only does she know the principles of leverage, but she also knows that sin(θ) + cos(θ) can be greater than 1.
Her best strategy is to lift and push at the end of the log (probably at ~15 degrees -- but I'll do the math later). This has the added advantage of kicking the axis of rotation down past the center of mass. ... (to be continued)
Lifting at θ = asin(1/10) gives you a little advantage. You can move a log with a k = 2cosθ/(10-sinθ) = 0.201 ... but that is not enough!
********
Dr. D has the right idea. Assuming for some reason that you could not get the instaneous axis of rotation to be offset from the center with a straight horizontal push as he proposes (I need to check his approach), you could with an angled upwards push. (It is this offsetting of the center of rotation which allows you to reduce the effective friction).
........
Dr. D appears to be right on spec. It also works with μ = 0.23. Note: the drat would be able to accelerate a little better with a very slight upward push -- it reduces friction and pushes the center of rotation a bit closer to optimum -- but the benefits are marginal.