I don't know how that would work, but it sounds like a reasonable idea if it's practical.
That prompts me to wonder the following: how do the outages occur? Is it from lack of enough power being generated at the time of peak demand, or is it because the mathematicians have forecast the right demand, and the grid - i.e. after the generation - gets fried by so much power being put into or drawn from it? I would have thought there would be preventative measures against this scenario.
If it's the former, then Ian's (latest) idea would have great merit indeed.
And if it's the latter, and the mathematicians have stuffed it up on one occasion and asked for a huge load of power which is not taken up by the consumers, can we still have an outage even though the demand was not that high? This might perhaps be brought about by a sudden and unforeseen change in the weather (prolly fairly unlikely I guess).