Books written by Ray Sullivan

Sunday, 15 December 2013

Sim Universe - The Ultimate Sim City Yet

The first time I played Sim City I was sucked in big time. The game was fairly new and relatively uncomplicated back then in the mid 90s. I upgraded to a faster computer (Pentium - wow) and picked up a copy of Sim City 2000 sometime in the early days of this Millennium and realised how much the game had developed and how little time I had for frittering my life away. Plus the game designers clearly were targetting a new audience with The Sims turning up, allowing you to custom build people. Consequently I haven't touched Sim City in any form for a decade, however I do have fond memories of building my first metropolis or two. I'm still a little miffed by the appearance of Godzilla who randomly destroyed what was my best city ever. I still feel that was uncalled for.

There's something about building a simulation that is satisfying, although if you dig a little you start to realise how difficult it is to model the real world with any accuracy. If it was easy then there would be no need for surprises in life and we'd be able to predict earthquakes, for example, with some accuracy. It would also negate the need to engage in pointless discussions regarding the probability of snow on Christmas Day, which would be a boon. It won't here, in case you're up for a debate. That would be pointless. But back to simulations.

In 2003 a philosopher named Nick Bostrop mooted the idea that we might potentially be part of a master computer simulation, a la Matrix, and challenged whether we would know if we were. It is a fair metaphysical question, up there with the one about trees falling in the forest with nobody to witness the event - did it make sound? The real question is, who cares? As is the way with philosophers that was about as far as Nick got to answering his question, but it probably generated a fair old discussion in college philosophy classes afterwards. One person who took the question seriously, though, was Professor Martin Savage, a physics researcher at the University of Washington.

Martin has used the metaphysical question to ask the real question - how could we tell? It doesn't mean he believes there is a real possibility that we are actually part of a computer simulation, but more of how the task could be tackled and what compromises would need to be taken in the process. You see, our planet and universe is an extraordinarily complex system, from the detail you see around you to the red shift of the big bang. Some parts of the known universe require literally the use of infinite parameters. Difficult for us lay people to get a handle on, even trickier for a computer given their distinctly finite resources. I've had the pleasure recently of reading a book, Quantum by Manjit Kumar, that took me through the discovery and development of quantum mechanics from Planck's Constant, Einstein's solution to Brownian Motion and Special Relativity, through to the developments and debates that ripped classical mechanics apart and established quantum mechanics as the current best interpretation of how the universe ticks. It didn't make me a physicist, nor did it bamboozle me with arcane descriptions and formulae, but instead placed me in the centre of the debates, with the many back stories that happened along the way. Well recommended for anyone with an interest in the subject that doesn't want to be daunted by the concepts.

Having had a recent primer, even one that didn't get as far as discussing lepton and quarks, helped me as I read through the lay presentation on Martin's team's work. Not essential, but extremely helpful as the presentation discusses lattice spaces, the likely computational capabilities over the next 50 years and the probabilistic qualities of quantum physics. As an example, quantum mechanics necessarily has to consider all the potential occurrences in a given spacetime frame before determining that the most likely outcome is the one that classical mechanics would generally determine, at least at a local level, but acknowledging that there is a remote but calculable probability that a different result may occur. That'll be Godzilla razing my city to the ground, then. The upshot is that a simulation could be constructed and a sentient race hell bent on annihilating itself sooner or later (sound like any sentient race you know of?) may well produce such a simulation, perhaps to perpetuate the race after doomsday. Or is that my SciFi plot generator working overtime again?

However the need to simulate infinite space, for example, means that some parts of the simulation will need to be fudged, made to look infinite using finite computational resources. It is the nature of this fudge, or signature as Martin puts it, that the research has looked at. I guess that if you can spot the signature behind what would be the most elaborate simulation possible, the spin off is likely to be a toolkit to analyse less comprehensive simulations, perhaps exposing their inherent weaknesses to allow better models to be constructed. So more warning about earthquakes, less chat about snow on Christmas Day and hopefully my next Sim City won't be trashed by Godzilla.

If you have the slightest interest in quantum mechanics you may want to look at Martin's presentation and read Manjit's book.

Visit my Book Website here
        Visit Project: Evil Website here                                        Visit DLF Website here

        Follow me on Twitter  - @RayASullivan

        Join me on Facebook -  use to find me

No comments:

Post a Comment