COMPUTING THE UNIVERSE

Immense simulations model billions ofyears of cosmic evolution

Dark, weblike patterns appear on the screen. Separate strands tangle and merge,
forming a blob that writhes and contorts like an octopus. Some fragments fly off
from this pulsating creature; others are drawn into the fiery blaze at its
center. The special effects, although impressive, are no threat to the likes of
Steven Spielberg. But this picture isn't intended to be entertaining; rather,
its ``director,'' MIT physicist Edmund Bertschinger, wants to model the
formation of the universe.

Bertschinger's models start about 10 million years after the Big Bang and run
forward in time to the present era. All the while, the computer tracks the
meanderings of 23 million particles. Even though the interactions are based on
simple laws of gravity worked out by Isaac Newton three centuries ago, the
resultant motions are complex and often counterintuitive.

Bertschinger and many others are focusing on one of the central problems of
cosmology: how, in the 10 to 20 billion years since the Big Bang, matter came to
arrange itself in the patterns seen today--vast sheets and strings of galaxies,
galaxy clusters, and clusters of clusters, separated by giant voids. Despite
all that they've learned, astronomers still don't have a clue as to what type of
matter makes up 90 to 99 percent of the stuff in the universe. All this hidden
material, which provides the glue holding galaxies and larger celestial
structures together, remains unseen and is thus ``dark matter.'' The models that
theorists play with make various guesses as to what this invisible matter might
consist of.

The most successful model to date attributes the formation of large-scale
structures to a class of unidentified flying particles called ``cold dark
matter'' (CDM). This model has come under fire in recent years as astronomers
have found ever-larger cosmic entities that CDM theory has trouble explaining.
However, the calculations of Bertschinger and a former graduate student, James
Gelb, show that CDM may, in fact, be adjusted to account for these mammoth
conglomerates. But then individual galaxies--about the smallest things in the
simulations--become too massive. The picture doesn't turn out right on both the
largest and smallest scales.

Cosmologists wouldn't have spotted this problem without computer simulations,
explains University of Toronto astrophysicist Nick Kaiser, because the CDM model
was only off by a factor of two. The next step is to experiment with other types
of dark matter until the picture the computer spits out is consistent with that
produced by astronomers diligently mapping the heavens.

The alternative model that has Kaiser and Bertschinger most excited is based on
``mixed dark matter''--that is, ``hot'' (or fast-moving) particles (most likely
neutrinos) as well as the slower-moving cold dark matter. So far, the
simulations of a mixed-dark-matter universe look pretty good.

Don't write off CDM yet, argues University of Chicago cosmologist Michael
Turner, although he concedes that CDM theory may indeed have serious problems.
On the other hand, the problem may lie with the simulations themselves. After
all, he adds, ``simulating the universe is a very tricky business.''

While Bertschinger acknowledges that there are limits to what we can glean from
simulations alone, he's confident that science will eventually sort things out.
``Although I'm pessimistic about CDM, I'm optimistic by and large because we're
learning from these simulations. They're teaching us some new physics that can
eventually guide us to new models.