Keep on rollin’

Attracted by algorithms some while back, I bought myself this audiobook… it sounded like exactly what I was looking for:

I found it hard going if I’m honest. Listening at 1.25x speed helped.

But while the ‘optimisation problems’ of public parking turned out to be vaguely useful at work, my heart sank a bit the other Saturday when rolling the ‘relevant complexity’ dice served up the answer: ‘Audiobook’…

…Ho hum. But you have to trust the dice.

And sure enough ‘Algorithms to live by’ proved the point; the chapter on ‘Randomness’ validates a lot of what I’ve been trying recently:

Recent work in computer science has shown that there are cases where randomized algorithms can produce good approximate answers to difficult questions faster than all known deterministic algorithms.

One problem they help with is ‘Hill climbing’ and local maxima.

At any point in life – however much you’ve perfected it, the risk is it could still be better. Like a climber in the mist, you know you’re in a good place – but there might be an even better one you can’t see for the fog.

And randomness is the way to find out.

Try something a bit different and you can find out whether you’re at the top of a small hill of possibilities – or surveying the entire range from the highest vantage point.

The truth of life is: you never can know if you’re stuck in a local maximum. But the odd throw of the dice now has the full weight of computer science behind it!

Keep on rollin!