Professional Documents
Culture Documents
Imagine trying to describe the circulation and temperatures across the vast
expanse of our oceans. Good models of our oceans not only benefit fish-
ermen on our coasts but farmers inland as well. Until recently, there were
neither adequate tools nor enough data to construct models. Now with
new data and new mathematics, short-range climate forecasting—for
example, of an upcoming El Niño—is possible.
w w w. a m s . o r g / m a t h m o m e n t s
MM/1
Making Movies
Come Alive
Many movie animation techniques are based on mathematics. Characters,
background, and motion are all created using software that combines pixels
into geometric shapes which are stored and manipulated using the mathe-
matics of computer graphics.
Software encodes features that are important to the eye, like position,
motion, color, and texture, into each pixel. The software uses vectors,
matrices, and polygonal approximations to curved surfaces to determine the
shade of each pixel. Each frame in a computer-generated film has over two
million pixels and can have over forty million polygons. The tremendous
number of calculations involved makes computers necessary, but without
mathematics the computers wouldn’t know what to calculate. Said one
animator, “. . . it’s all controlled by math . . . all those little X,Y’s, and Z’s that
you had in school—oh my gosh, suddenly they all apply.”
w w w. a m s . o r g / m a t h m o m e n t s
MM/2
Designing Aircraft
The flow of air (and water) has been studied for over a hundred years, but
only recently have mathematicians begun to understand the complicated
phenomenon of turbulence that is a crucial part of aerodynamics. With
mathematics and modern computers, wind tunnels are now seldom used in
aeronautical design.
w w w. a m s . o r g / m a t h m o m e n t s
MM/3
Securing Internet
Communication
No one could shop, pay bills, or conduct business securely on the Internet
without the mathematics of encryption. Although based on algebraic facts
proved centuries ago, today’s sophisticated encryption techniques were
formulated within the past twenty-five years.
Public key encryption allows a user to publish the encryption key for all to
use, while keeping the decryption key secret. One such algorithm, called
RSA, is behind the encryption in modern browsers. The National Institute
of Standards and Technology recently adopted an Advanced Encryption
Standard that will be used for electronic communication in the years to
come. This new standard uses permutations, modular arithmetic, polyno-
mials, matrices, and finite fields to transmit information freely but securely.
w w w. a m s . o r g / m a t h m o m e n t s
MM/4
Deciphering DNA
Anyone who has used a garden hose knows that knots appear in strange
places. Scientists have found that a branch of mathematics called knot
theory appears in many familiar places, including in our DNA. Mathematics
plays a key role in understanding how DNA functions and replicates itself.
Certain enzymes cut a strand of DNA at one point, pass another part of the
strand through the gap, and then seal the cut. Knot theory gives insight on
how frequently an enzyme has to act, from which one can infer how long
the enzyme might take to make a product. This kind of complex manipula-
tion is significant in many cellular processes—including DNA repair and gene
regulation—and is the type of problem central to the theory of knots.
w w w. a m s . o r g / m a t h m o m e n t s
MM/5
Listening to Music
No matter how complicated the music (or data), from Mozart to Twisted
Sister, it is stored on disks using only the numbers 0 and 1. To do this, many
different branches of mathematics, both advanced and elementary, are used
at each step of the process.
Signal Processing: The original sound is sampled, measuring the sound waves
at regular, frequent intervals. How frequently depends on the Shannon
Sampling Theorem.
Binary Arithmetic: The amplitudes are represented as a sixteen-bit sequence of
0’s and 1’s. The 0’s and 1’s are stored on the CD as smooth areas and pits.
Partial Differential Equations: Equations in fluid dynamics govern the process
of compressing the reflective and protective layers over the data.
Linear Algebra: Inevitable corruptions of the 0’s and 1’s (dust or scratches,
for example) are compensated for with error-correcting codes.
Trigonometry and Calculus: To retrieve the data, a tracker moves a laser
which is focused on the data. As the laser reads from the center of the disk
to its edge, a motor must continually move the CD slower to keep the
speed of the data reading constant.
For More Information:
Scientific American, Ken C. Pohlmann, 1998.
w w w. a m s . o r g / m a t h m o m e n t s
MM/6
Forecasting Weather
Forecasting the weather requires enormous amounts of data and computa-
tion. In order to have an accurate model of the weather, one must know
the temperature, humidity, air pressure and wind speed (among other things)
at different points and elevations. Although incorrect forecasts may be more
memorable, current three- to seven-day forecasts are better than 36-hour
forecasts were just 20 years ago. Increases in computing power have helped
improve weather forecasts, but it’s the mathematics behind the models that
has led to the great increase in accuracy.
w w w. a m s . o r g / m a t h m o m e n t s
MM/7
Storing Fingerprints
Storing and identifying the digitized version of millions of fingerprints is an
almost inconceivably enormous task. Uncompressed, the FBI’s current
fingerprint files would consist of 200 terabytes (200,000,000,000,000 bytes).
A new piece of mathematics, wavelets, makes data compression fast, rela-
tively routine, and much less expensive so that storage is feasible and
retrieval is fast.
Any image is really a function that gives the color and intensity of each pixel.
This function can be written as a combination of special functions—the
wavelets. The rules for how the wavelets fit together are easier to store
and retrieve than the function itself. Wavelets are a twofold improvement
over Fourier transforms—another data compression technique based on
sines and cosines.
w w w. a m s . o r g / m a t h m o m e n t s
MM/8
Creating Crystals
Both the power of mathematics and the speed of today’s computers are
needed in the study of crystal formation. In addition to the aesthetically
appealing possibility of understanding snowflakes, the field of crystal forma-
tion is crucial to the integrity of steel, superconductors, and computer chips.
While crystals are forming, they have moving, irregularly shaped boundaries,
which makes only numerical solutions to their equations possible. Part of
crystal formation follows the principle of a minimum surface area for a fixed
volume, but the orientation of a crystal also affects its formation: Heat is
diffused more easily away from the surface than into it, so crystals in the
direction of the exterior form faster than others. The extra complexity that
orientation brings to the problem of crystal formation makes solving the
relevant equations more difficult.
w w w. a m s . o r g / m a t h m o m e n t s
MM/9
Experimenting with
the Heart
Experimenting with real human hearts isn’t possible, but experimenting with
accurate mathematical models of the human heart has led to a new under-
standing of its complex processes. Mathematics and the computer can
replace years of experimentation in laboratories. For example, under-
standing resulting from mathematics greatly speeds up the design and
implementation of artificial valves.
Equations based on Hooke’s Law model the geometry of the heart by repre-
senting muscle fibers as closed curves of different elasticities. The Navier-Stokes
equations, which describe all fluid flows, model blood flow in and around the
heart. The fact that the heart’s shape is constantly changing, however, makes the
equations especially hard to solve, and a precise solution to the equation can’t
be found. Approximate solutions are generated by computer.
w w w. a m s . o r g / m a t h m o m e n t s
MM/10
Seeing the World
through Fractals
Fractals are self-similar mathematical objects that make computer graphics
and simulations more realistic. The self-similarity of fractals is like that of a
fern or a country’s coast: successive magnifications yield images, each one
resembling the original.
w w w. a m s . o r g / m a t h m o m e n t s
MM/11
Tracking Products
Number theory, an area of mathematics on which the security of Internet
communication is based, also guarantees the authenticity of a book’s ISBN
(International Standard Book Number) and its UPC (Universal Product
Code). Authentic ISBN numbers allow computerized tracking of books, and
UPC’s identify each book as unique—ensuring that authors get the royalties
they’re due.
All digits of an ISBN number but the last are multiplied in modular arith-
metic by predetermined numbers. The sum of these products is used to
create a check digit. The check digit is then appended to the end of the
number. Not only does the check digit guard against intentional fraud, but it
also detects human error, like transposing two digits. This process is also
used with most state driver’s licenses and with airline tickets.
More information:
Contemporary Abstract Algebra, Joseph A. Gallian, 1998.
w w w. a m s . o r g / m a t h m o m e n t s
MM/12
Investing in Markets
The past twenty years has seen the creation of many new sophisticated
financial instruments, such as derivatives, which have helped drive the
economy. Financial derivatives are mathematical instruments whose value is
derived from the value of something else, and while some view them as
risky, their purpose is to lessen risk by sharing it with others.
w w w. a m s . o r g / m a t h m o m e n t s
MM/13
Routing Traffic
through the Internet
Understanding the way packets of information move through the Internet is
a challenging problem. Internet traffic behaves quite differently from tradi-
tional phone-line traffic. Fractal-based modeling has been successful in
describing aspects of Internet data traffic ranging from the inter-keystroke
times of a person typing to the sizes of files transferred.
w w w. a m s . o r g / m a t h m o m e n t s
MM/14
Manufacturing
Better Lenses
The design of eyeglass lenses, especially for progressive prescriptions, is a
part of mathematics which involves geometry, materials science, and partial
differential equations in a surprising way. It is an active area of research that
affects people’s lives every day (especially those of us over forty).
w w w. a m s . o r g / m a t h m o m e n t s
MM/15
Mapping the Brain
Mathematics is used to understand how to precisely identify the parts of the brain
that correspond to specific functions. Current research involves mapping our three-
dimensional brain to two dimensions, similar to translating a globe to a map. Yet
because of the many fissures and folds in the surface of the brain, mapping our
brains is more complex than converting a globe to a map.
Points of the brain that are at different depths can appear close in a conventional
image. To develop maps of the brain that distinguish such points, researchers use
topology and geometry, including hyperbolic and spherical geometry. Conformal
mappings — correspondences between the brain and its flat map that don’t distort
angles between points — are especially important to accurate representations of
the brain. Just as a map of the earth aids navigation, conformal mappings serve as
a guide for researchers in their quest to understand the brain.
w w w. a m s . o r g / m a t h m o m e n t s
MM/16
Eye-dentifying Yourself
Iris recognition may allow us to live in a world without PIN numbers—iden-
tifying ourselves just by looking at the ATM. Identification by iris recognition
is based on pattern recognition, wavelets and statistics.The first two fields
are used to translate the patterns in your iris into a string of 0’s and 1’s,
while statistics establishes that the scanned iris is yours.
The iris is a good physical feature to use for identification because of the
tremendous variability in iris patterns, even between twins.This variability
guarantees that a correct identification is made when the code for a
scanned iris matches a stored code in at least two-thirds of the bits.
Furthermore, the eye and iris are easy for a scanner to find, due to their
shape and placement. Once the iris is located, wavelets are used to translate
the pattern of the sampled portion of the iris into two bits.These bits
reflect the agreement between that portion of the iris and specific wavelets.
The entire iris is encoded in about 2000 bits. Finding a relative match
between this bit pattern and one of the thousands of iris codes in the data-
base completes the identification.This comparison is done in parallel, so that
the whole process takes place in about the blink of an eye.
w w w. a m s . o r g / m a t h m o m e n t s
MM/17
Bidding Wisely
The basic auction format familiar to most people is just one of the many
forms that an auction might take. For example, in the "second-price auction,"
the highest bidder is awarded the object but charged only the amount of the
second-highest bid. (This is not as bad for the seller as it might sound; bidding
can be more aggressive in this format.) This and other auction formats are
studied using mathematical modeling, game theory, combinatorics, integer
programming and optimization. One fundamental conclusion arrived at by
researchers is that inexperienced bidders almost always overbid.
The Internet is one factor that has led to an increase in the number of
items sold at auction. Reverse auctions—in which a company needing a
product allows providers to bid on the price, with the lowest bidder
winning—are also gaining in popularity. Some newly designed auctions allow
bids on groups of items or involve more than one round of bidding.The
increased frequency of auctions is exemplified in the airline industry, where
the gate a flight leaves from and the amount of the voucher you receive for
surrendering your seat may both be determined by auction.
For more information: The Economic Theory of Auctions, Paul Klemperer, ed.
Dana Breslin/Art270
w w w. a m s . o r g / m a t h m o m e n t s
MM/18
Making Votes Count
The outcome of elections that offer more than two alternatives but with
no preference by a majority, is determined more by the voting procedure
used than by the votes themselves. Mathematicians have shown that in such
elections, illogical results are more likely than not. For example, the majority
of this group want to go to a warm place, but the South Pole is the group’s
plurality winner. So if these people choose their group’s vacation destination
in the same way most elections are conducted, they will all go to the South
Pole and six people will be disappointed, if not frostbitten.
Elections in which only the top preference of each voter is counted are
equivalent to a school choosing its best student based only on the number
of A’s earned. The inequity of such a situation has led to the development
of other voting methods. In one method, points are assigned to choices, just
as they are to grades. Using this procedure, these people will vacation in a
warm place—a more desirable conclusion for the group. Mathematicians
study voting methods in hopes of finding equitable procedures, so that no
one is unfairly left out in the cold.
w w w. a m s . o r g / m a t h m o m e n t s
MM/19
Simulating Galaxies
Galaxies can be more than 100,000 light years across, consisting of hundreds of
billions of celestial bodies, and with a mass more than a trillion times that of our
sun. Modeling such huge, complex systems, in which many of the stars have
chaotic orbits, requires new computational techniques. Advances in the speed
and memory of computers have improved models, as has parallel computing, but
advances in algorithms—the way the mathematics of a problem is converted
into steps a computer can perform—are indispensable in developing accurate
galaxy models.
The complexity of simulating the behavior of a galaxy is not limited to the galaxy
itself. Since a galaxy is usually part of a cluster or supercluster of galaxies, the
external forces exerted by these larger agglomerations on the galaxy must also be
accounted for.Thus, models must be accurate across many scales of distance. Instead
of numerically solving the equations of the model uniformly across all sectors,
researchers employ multi-scale algorithms that do more calculations in sectors
determined to be more significant.This kind of technique uses computing power
more efficiently, giving us a glimpse of the underlying structure of the universe.
w w w. a m s . o r g / m a t h m o m e n t s
MM/20
Revealing Nature’s
Secrets
Mathematical ecology is a growing and active area of interdisciplinary research
between mathematics and ecology, using almost every part of mathematics (linear
algebra, analysis, differential equations, stochastic processes, numerical simulations,
statistics) to understand and model complex biosystems.This modeling helps
establish important parameters and thresholds, such as the area required to
sustain a species or how fast an invasive species will spread through a region.
Models must be fairly complex to capture how a single species interacts with
other species and with its environment.Today’s mathematical ecology
researchers are faced with the far more daunting task of simulating several inter-
connected networks of organisms across different scales of time, size, and space.
To do that, researchers resort to some relatively new areas of mathematics, for
example non-linear dynamical systems and spatial statistics.
w w w. a m s . o r g / m a t h m o m e n t s
MM/21
Folding for Fun
and Function
Origami—paper-folding—may not seem like a subject for mathematical investi-
gation or one with sophisticated applications, yet anyone who has tried to fold a
road map or wrap a present knows that origami is no trivial matter.
Mathematicians, computer scientists, and engineers have recently discovered that
this centuries-old subject can be used to solve many modern problems.The
methods of origami are now used to fold objects such as automobile air bags
and huge space telescopes efficiently, and may be related to how proteins fold.
w w w. a m s . o r g / m a t h m o m e n t s
MM/22
Getting Results
on the Web
Imagine trying to find the right information quickly in a library where billions of
pages are randomly piled in a heap, instead of being in books shelved in order.
That's what Web search engines do, millions of times a day. First-generation search
engines often found useful pages, but those pages may have been too far down the
list to be of any practical use. Current search engines rank pages by using mathe-
matics—probability, graph theory, and linear algebra—so that sites most relevant
to a query are listed at the top, where the user can most easily see them.
The vast number of pages and links on the Web can be represented as a graph in
which the nodes are Web pages and the directed edges are links.Today's search
engines determine the relevance of a page to a query by incorporating the impor-
tance of pages pointing to and from that page.Thus, when it comes to a search, a
page’s links can be just as important as its content.The final ranking comes from
techniques in linear algebra and probability that help formulate and solve equa-
tions which, according to the founders of one search engine, involve millions of
variables and billions of terms. In the future, search engines may use artificial intel-
ligence and information on past searches to discern the actual intent of a query.
w w w. a m s . o r g / m a t h m o m e n t s
MM/23
Enhancing Your Image
Inpainting is the age-old practice of restoring visual works— a process that until
recently was only performed manually by experts. Many people now use
computers to retouch digital photographs, yet the work can still be painstaking.
A promising new field of mathematical research is the development of
algorithms that solve partial differential equations to digitally inpaint with little
input or effort from the user. The technique can also be used, as in the example
below, to recover missing portions of transmitted images without requiring
retransmission of the data.
The apparent ease with which these new algorithms restore pictures masks the
difficulty of creating software to imitate the trained eye and hand of a
professional. Digital inpainting methods must incorporate information not only
about colors near the incomplete area, but also about the direction of change in
the boundaries between existing lines and missing ones. Some inpainting
procedures rely on techniques from computational fluid dynamics, ensuring that
known information “flows” continuously into needed areas. Thus, results from
the well-established field of computational fluid dynamics are applied to the new
field of digital inpainting so that everyone can get the complete picture.
For more information: “Filling in Blanks,” Ivars Peterson, Science News,
11 May 2002.
w w w. a m s . o r g / m a t h m o m e n t s
MM/24
Defeating Disease
From modeling microscopic genes and proteins to tracing the progression of an
epidemic through a country, mathematics plays an important role in combating
disease. For example, the basic model used to analyze the dynamics of infectious
disease is a system of differential equations. A new field called “data mining”, involving
statistics and pattern recognition, helps locate significant information in the vast
amounts of data collected from studies of diseases in populations. Mathematics also
plays a key role in connecting changes in the human genome to specific diseases.
Mathematics has helped recent fights against foot-and-mouth disease in the United
Kingdom and against Chagas disease — a disease affecting millions of people in Latin
America. Epidemiologists studying the foot-and-mouth epidemic used mathematical
models to conclude that early efforts were insufficient to stop what would become a
calamitous spread of the disease. The government accepted the conclusions and took
a course of action that, although drastic, did indeed arrest the outbreak. In Latin
America, mathematicians computationally tested several courses of action against
Chagas disease and found a surprisingly simple yet highly effective step (keeping dogs
out of the bedroom) to greatly reduce the infection rate. These examples share three
important characteristics: a mathematical model of the disease, modern computers to
do calculations required by the model, and researchers with the insight to design the
former so as to take advantage of the power of the latter.
For more information: Infectious Diseases of Humans: Dynamics and Control,
R. M. Anderson and R. M. May.
protein, enhances display of icosahedral symmetry.
Wisconsin-Madison. Rhinovirus color-coded by
Image courtesy of: Jean-Yves Sgro, University of
©1993
w w w. a m s . o r g / m a t h m o m e n t s
MM/25
Revolutionizing Computing
In about 20 years, computer chips will be so small that the effects of quantum
mechanics will replace the physical laws we take for granted. While today’s
computing is based on bits that are either 0 or 1, the basic unit in quantum
computing is the quantum bit—the qubit—which can be 0 and 1 simultaneously
(with a probability associated with each). In the strange world of quantum
computing, complicated procedures such as factoring large numbers are done much
faster because the many steps involved can be done concurrently.The ultimate goal
of mathematicians, physicists, computer scientists, and engineers in the field is to
create a quantum computer that could solve in seconds some problems that would
take today’s most powerful computers billions of years to solve.
Among the capabilities of a quantum computer would be the ability to do the calcu-
lations necessary to break today’s electronic encryption methods. This is not as
alarming as it may sound, because cryptographers have already designed algorithms
to take advantage of the quantum mechanics principle that observing a system’s
state changes it.Thus, users of a quantum communications network could detect any
attempt to intercept their communication. It is somehow ironic that the laws that
govern the barrier to the miniaturization of today’s computers may provide a boon
to future computing.
For more information: “Rules for a Complex Quantum World,” Scientific American,
November 2002, Michael A. Nielsen.
Image courtesy of the MITRE Corporation.
w w w. a m s . o r g / m a t h m o m e n t s
MM/26
Cutting the Cord
A cellular phone’s size disguises the considerable amount of activity going on inside.
In a digital phone, your voice is converted by the phone’s processor into a stream
of 0’s and 1’s that are transmitted to a base station, received, relayed, and recon-
verted back to the original sound (actually, an extremely good approximation of
that sound) by the receiving phone. Along with sending your words, your phone
transmits an identifying code and determines the nearest base station. Hand-off
algorithms are employed to help maintain a continuous conversation as the phone’s
location changes. (Note that E.T. didn’t phone home until after landing.)
w w w. a m s . o r g / m a t h m o m e n t s
MM/27
Expressing Yourself
The state-of-the-art technology used by researchers to identify active
(expressed) genes in cells is the microarray: a “gene chip” imprinted, not with
circuits, but with DNA. Active genes of fluorescently tagged cell samples placed
on the chip reveal themselves when they bind with their DNA complements on
the chip.The amount of data generated by this microscopic activity is enormous:
just one row in an array can have 15,000 points. Pattern recognition and image
analysis are two fields which use mathematics to help extract important genetic
information about several diseases, including Alzheimer’s and Parkinson’s, from
microarray data. In the future, microarrays may enable an individualized
approach to medicine, in which your doctor could use these chips to diagnose
disease and determine the best treatment for your unique genetic profile.
In one particular area of medicine, cancer research, the points in each column of
an array can be thought of as genetic coordinates of samples from tumors.Yet
there are so many coordinates that it is difficult to determine which tumors are
similar. Algorithms employ statistics and different measures of distance in higher
dimensions to group genetically similar
tumors into “clusters” so that experi-
ments can be done on treatments
corresponding to the clusters. In one
case, microarray technology not only
Clive and Vera Ramciotti Functional Genome Array Centre.
w w w. a m s . o r g / m a t h m o m e n t s
MM/28
Bringing Robots to Life
Robots of all shapes and sizes now perform tasks as routine as vacuuming the
living room floor and as remarkable as discovering a hydrothermal vent on the
ocean floor. Geometry, statistics, graph theory, differential equations, and linear
algebra are some of the areas of mathematics that allow navigation and decision
making so that robots can function autonomously and do things we either can’t,
or would rather not, do.
The robot pictured below not only dances but also greets visitors and escorts
them to their destinations, providing news and weather updates along the way.
Abilities like these require algorithms for vision, pattern recognition, speech
recognition, and dealing with uncertainty so that accumulated error doesn’t
render the robot ineffective. Most researchers think that we are a long way
from creating machines that behave like humans, but improving algorithms will
improve the capabilities of robots, which have already served in space, in rescues
at disaster areas, and in the operating room, where physicians use robotic arms
that allow for more precise, less invasive surgery.
Getty Images
w w w. a m s . o r g / m a t h m o m e n t s
MM/29
Making Connections
People in a society, neurons in the brain, and pages on the Web, along with their
connections, are all examples of networks. Mathematicians study characteristics of
networks, such as the number and distribution of connections, to discover what such
attributes may reveal about the intrinsic nature of a network. For example, the colors
in the picture below indicate how disruptive deleting a node would be to the
network, in this case a living cell.The discovery and verification of network properties
such as this has significance for applications ranging from the microscopic to the
worldwide, including the protection of both computers and humans against viruses.
The study of networks spawned the phrase “six degrees of separation”, the theme of
a game involving actors’ connections via common film appearances. In an experiment
done in the 1960s, over 100 randomly chosen people in the Midwest were found to
be connected to a Massachusetts stockbroker (by a friend of a friend of a friend, and
so on) in an average of just six steps.That people halfway across the country could
be so closely connected was quite a revelation and proved that even a large network
could be a “small world”.Today, researchers use parameters from graph theory and
probability in analyzing networks to determine whether an elaborate network, be it
a power grid or actors connecting to Kevin Bacon, is indeed a small world after all.
w w w. a m s . o r g / m a t h m o m e n t s
MM/30
Beating Traffic
It’s not your imagination; traffic is getting worse. In the last 30 years while the
number of vehicle-miles traveled has more than doubled, physical road space has
increased only six percent.Yet building new roads is no guarantee of relief:
A counterintuitive result in traffic science is that a new road could actually
increase the congestion in a network. Areas of mathematics like queuing theory
and partial differential equations contribute to understanding traffic, which is a
backwards propagating wave —cars move forward but the jam moves backward.
The mathematical study of traffic is relatively new, but a federal report concluded
that the information revolution—that is, the combination of more powerful
computers, telecommunications, and better numerical models—will affect
transportation as much as the inventions of the automobile and jet engine.
Analyzing traffic (like predicting weather ) requires many variables (driver speed,
length of trip, time of day, and origination point) and involves chaos theory (a small
change down the road can drastically change travel conditions). Unlike weather,
however, traffic can change in response to a forecast as alternative routes are
chosen—today by drivers and, in the future, perhaps by the cars themselves.
w w w. a m s . o r g / m a t h m o m e n t s
MM/31
Charging Through Space
Electromagnetic disturbances on the sun usually don’t affect us 93 million miles away,
but a major solar storm can, with severe consequences for satellites, electricity, and
communications. For example, in 1989 solar explosions resulted in the collapse of a
major power grid, leaving over six million people in Canada without power. Space
weather forecasters now have better mathematical models from which they make
statistical predictions about solar activity and its effects.The predictions have
improved with technology, but without new mathematics and a refinement of the
models, even the best computers would be lost in space.
Space weather models are based on Maxwell’s electricity and magnetism equations
and fluid flow equations, which due to their complexity, must be solved numerically.
Newly launched satellites—including four which maintain a tetrahedral formation and
provide a three-dimensional image of space weather—supply needed information to
improve our understanding of the space environment and enable warnings about
potential disruptions to modern amenities.
For More Information: Storms from the Sun, Michael J. Carlowicz and
Ramon E. Lopez
w w w. a m s . o r g / m a t h m o m e n t s
MM/32
Tracing Your Routes
The Traveling Salesman Problem entails finding the shortest route that passes through
each assigned town exactly once. (The route below visits over 13,000 towns.) The
problem is noteworthy for its complexity, which grows exponentially with the
number of towns, and for its applications, which range from wiring a chip to sched-
uling airline crews. Researchers use graph theory and linear programming to solve the
problem when feasible and to find near-optimal solutions in other instances, saving
industry time and money.
There may never be a workable general solution to the Traveling Salesman Problem.Yet
even without knowing the best answer, mathematicians still can estimate how close to
optimal a given route is. Perhaps even more surprising: Operating on a map of 25,000
towns, current algorithms design paths whose lengths are within 0.01% of that of a
shortest path.
w w w. a m s . o r g / m a t h m o m e n t s
MM/33
Unlocking the Cell
The processes that cells perform are as wondrous as their individual mechanisms
are mysterious. Molecular biologists and mathematicians are using models to begin
to understand operations such as cellular division, movement, and communication
(both within the cell and between cells).The analysis of cells requires many diverse
branches of mathematics since descriptions of cellular activity involve a combina-
tion of continuous models based on differential equations and discrete models
using subjects such as graph theory.
It may be surprising, but cell functions are depicted with complex wiring diagrams
of circuits with signaling pathways, gates, switches, and feedback loops. Researchers
translate the diagrams into equations, which are often solved numerically. Solving
the equations is only part of a process in which solutions are analyzed, models are
refined, and equations are reformulated and re-solved.This may be repeated many
times.The aim of this process is an accurate representation of cell behavior, which
may allow drugs and treatments to be designed in the same precise way that elec-
tronic circuits are today.
w w w. a m s . o r g / m a t h m o m e n t s
MM/34
Seeing More Clearly
Twinkling stars are fun for songs but frustrating for astronomers. Current tech-
nology uses adaptive optics to adjust for turbulence in the atmosphere and deliver
an accurate image of stars, planets, and satellites. Correcting for atmospheric
distortion involves linear algebra, geometry, and statistics to determine the extent
of the distortion and continually adjust deformable mirrors which refocus light
waves back along their true paths.
w w w. a m s . o r g / m a t h m o m e n t s
MM/35
Canning Spam
Email users ask the same question as surprised diners:Who ordered the spam?
The answer is no one, but sending back email spam only brings more. People are
fighting spam with many new tools, including filters that look for telltale signs that
a messsage is spam. Spammers, however, defeat simple filters by disguising the
words and intent of their messages. New, more sophisticated filters use mathe-
matics to fight spam by training the filters to recognize spam over time, so that
your server brings you what you want.
Spammers adapt their messages to avoid many anti-spam tools, but using a mathe-
matical result known as Bayes’ Theorem, the tools can adapt as well. As users
examine email each day, they indicate which messages passing through the filter
are, in fact, spam.With training, the filter learns how likely it is that certain words
or characteristics are present when a message is spam. Bayes’ Theorem allows the
filter to turn this information around, calculating how likely it is that the message is
spam when those words or characteristics are present. It is a powerful application
of an old and fundamental mathematical result. Using new and old mathematical
tools, mathematicians continue to work on innovative techniques to combat spam.
For More Information: “Math 1, Spam 0,” Dana Mackenzie, SIAM News,
November, 2003.
w w w. a m s . o r g / m a t h m o m e n t s
MM/36
Locating, locating, locating
Originally designed for military use, the Global Positioning System (GPS) now lets
boaters, drivers, and hikers pinpoint their location to within a few meters. Most of
GPS's functionality is derived from arithmetic, algebra, and geometry.The time it
takes for a signal to travel from a transmitting satellite to a GPS receiver estab-
lishes the distance between the two, which places the GPS user on an imaginary
sphere centered at the satellite. Similar calculations are done concurrently using
other satellites. Once corrections for the difference between satellite and receiver
clocks are made, the GPS user's location must be one of the points of intersection
of three spheres.
The basic principles of GPS are simple, but reducing error when using satellites
more than 10,000 miles away to calculate locations is not. Information theory
extracts reliable data from weak signals (which have less than a billionth of the
power of those received by your television) and mathematical models of the
atmosphere account for slight changes in speed as signals travel through different
layers on their way to earth. Differential GPS reduces error even further by using
land-based stationary receivers, whose precise positions are known. Eventually
real-time GPS will be so accurate — with errors on the order of inches — that it
will guide cars and allow planes to land in zero visibility.
w w w. a m s . o r g / m a t h m o m e n t s
MM/37
Targeting Tumors
Detection and treatment of cancer have progressed, but neither is as precise as
doctors would like. For example, tumors can change shape or location between
pre-operative diagnosis and treatment so that radiation is aimed at a target which
may have moved. Geometry, partial differential equations, and integer linear
programming are three areas of mathematics used to process data in real-time,
which allows doctors to inflict maximum damage to the tumor, with minimum
damage to healthy tissue.
For More Information: “Treatment Planning for Brachytherapy,” Eva Lee, et al,
Physics in Medicine and Biology, 1999.
Institute of Technology.
w w w. a m s . o r g / m a t h m o m e n t s
MM/38
Being a Better Sport
From designing uniforms with less drag to adjusting the angle at which an athlete
launches a javelin, mathematics helps improve sports performance. Differential
equations and vector analysis play important roles in determining optimum
mechanics in a sport, as does numerical analysis when equations can't be solved
exactly. Many fields of mathematics are providing legitimate tools that allow
athletes to use mind and body to go swifter and higher.
w w w. a m s . o r g / m a t h m o m e n t s
MM/39
Recognizing Speech
Current speech recognition systems perform fairly well in non-conversational
settings such as dictation or requests for directory assistance. Applications like
this may not appear impressive, but because of accents, inflections, and pauses,
even such simple situations require sophisticated techniques to transform
speech waveforms into words accurately. One of the most common techniques is
a mathematical tool known as a hidden Markov model, involving conditional probabili-
ties, which trains on candidate sounds so as to locate the best match for a given
input.
w w w. a m s . o r g / m a t h m o m e n t s
MM/40
Compressing Data
Through digitization, films that require 10,000 feet of tape now fit on a disk less
than five inches in diameter. An important part of digitization is data compression,
which involves converting a large file to a smaller version, from which the original
(or a close approximation) can be recreated. Linear algebra, probability, graph
theory and abstract algebra are among the areas of mathematics at the foundation
of various compression algorithms that make modern technologies such as DVDs,
HDTV and large databases, possible.
No one technique can fulfill the compression requirements of all media. For
example, wavelet compression—based on a fairly new mathematical tool—works
well with images and audio files, but not as well with text files.Yet regardless of the
application, compression algorithms use redundancy and relatedness in data to
make storage and transmission more efficient. Does compression work? U b t jdg.
w w w. a m s . o r g / m a t h m o m e n t s
MM/41
Putting Together the Pieces
Fitting just-broken pieces together is hard enough, but imagine how difficult it
is to do after thousands of years — and a few civilizations — have passed.
Archaeologists faced with hundreds of thousands of pieces at a site have turned to
mathematicians to help reassemble the fragments. The pieces are first digitally
scanned; then software uses geometry, combinatorics, and statistics to
reconstruct ancient artifacts, even when many pieces are missing.
w w w. a m s . o r g / m a t h m o m e n t s
MM/42
Scanning the Unseen
By sending low-dose X-rays at an object through a range of angles and measuring
the rays’ absorption, CAT (Computed Axial Tomography)-scans provide precise
images that conventional X-rays can’t. Multivariable calculus and a mathematical
tool known as the Radon transform —invented early in the 20th century—are
crucial to the efficient reconstruction of a three-dimensional image from the
information gleaned along the one-dimensional lines.That efficient reconstruction
allows for better images with less exposure to X-rays— benefiting doctors and
patients alike.
The same mathematical principles used in CAT-scans are also used in a field called
astrotomography, providing unprecedented resolution of binary stars and the
surfaces of rapidly rotating stars. In this application the rotation of a star or pair of
stars replaces the rotation of the scanning machine and positions and velocities are
found based on radiation detected from the star(s).Thus, mathematics discovered
long before CAT-scan technology enables detailed views from within the human
body to far beyond our solar system.
w w w. a m s . o r g / m a t h m o m e n t s
MM/43
Making Designs A Reality
The innovative design of the Sydney Opera House stymied builders for years
until they realized that all the project's specifications could be met with triangles
cut from the same sphere. Since all the pieces were of the same type and from a
surface with well-established geometrical properties, the requisite calculations
(such as determining structural forces) were simplified considerably and the dream
became a magnificent reality.
For More Information: "Mathematical Tour through the Sydney Opera House,"
The Mathematical Intelligencer, Joe Hammer, Fall 2004.
w w w. a m s . o r g / m a t h m o m e n t s
MM/44
Making a Splash
The interplay of water, light, and music in some modern fountains is magical to
behold, and mathematics is part of that magic. Geometry is used in the overall
design, mathematical modeling simulates the fluid-particle interactions, and
powerful algorithms drive the software that coordinates thousands of valves and
lights through the numerous sequences in a typical show.
The ability to make water act so precisely results from the use of laminar flow
streams where all particles move in parallel and at the same speed. A complex
mathematical analysis of fluid dynamics makes it possible for water to perform
feats such as climbing stairs or behaving like individual marbles.The result is both
wondrous and efficient: A four-foot column of water wouldn't fill a normal
drinking glass.
For More Information: “Inventive Artist Sculpts in Water,” USA Today, Bill
Meyers, March 14, 1999.
w w w. a m s . o r g / m a t h m o m e n t s
MM/45
Reading Your Mind?
How does something the size of a yo-yo successfully play a game of 20 Questions?
Although its success tempts players to think that the device is reading their minds,
it's not.This sophisticated toy uses mathematics such as probability and fuzzy logic,
and mathematical objects such as matrices to determine your animal, vegetable or
mineral more than 75% of the time.
w w w. a m s . o r g / m a t h m o m e n t s
MM/46
Packing It In
Packing items into bins of given capacities may not sound important (unless you're
packing for a trip), but the topic of bin packing includes situations such as allocating
blocks of computer memory and scheduling airline flights as well as traditional
problems like loading trucks. Researchers use areas of mathematics (such as
number theory, geometry, and probability) to solve packing problems so that time
and storage – both physical and electronic – can be used efficiently.
Mathematicians proved that bin packing problems are “complex,” and a practical
algorithm that gives an optimal solution to all packing problems appears unlikely.
Yet even though there may never be a “fast” general solution, mathematicians still
seek to improve packing algorithms, saving industry time and money. One such
result demonstrates that one of the simplest packing algorithms, first loading the
biggest things that fit, is always within about 20% of the best solution possible.
w w w. a m s . o r g / m a t h m o m e n t s
MM/47
Translating: From Arabic
to Zulu
The current pace of document creation (on the Internet, for example) is much
greater than the capacity of human translators, which makes machine translation a
necessity. Machine translators use probability, statistics and graph theory in combi-
nation with databases of hundreds of millions of words and phrases in many
languages to achieve good translations efficiently.Thus, mathematics, often called
the universal language, also forms a bridge between languages.
Once a document is translated, the question becomes: How good is the translation?
Numerical measures of effectiveness help automate this part of the process as
well, saving time and money. Results from the evaluation improve translation
algorithms so that the urban legend of a computer translating “The spirit is willing
but the flesh is weak” into Russian and back into English as “The vodka is good but
the meat is rotten” will remain a legend.
For More Information: “Machine Translation in the Year 2004,” Kevin Knight and
Daniel Marcu, http://www.isi.edu/~marcu/papers/mt-icassp2005.pdf.
w w w. a m s . o r g / m a t h m o m e n t s
MM/48
Boldly Going
The “tubes” below are illustrations of low-energy pathways along which space vehicles
can travel using far less fuel.The recent discovery of these pathways has made previ-
ously impossible missions feasible. Much of space travel depends on calculus,
trigonometry, and vector analysis, but the existence of these routes derives from an
area of mathematics called dynamical systems applied to the mutual interaction of the
gravities of the sun, nearby planets, and moons.
Calculations of forces between two celestial bodies and their orbits are fairly direct,
but to understand orbits and trajectories when more than two bodies are involved,
dynamical systems and chaos theory are necessary. Even the simplest extension
beyond two bodies, the three-body problem, has been proven to have no explicit general
solution. Some special cases, however, have been solved and applied not only to
mission design, but also now to atomic physics to study the paths of certain excited
electrons.Thus, mathematics is locating new routes for space travel and establishing
connections between the atomic and the cosmic.
For More Information: "Ground Control to Niels Bohr: Exploring Outer Space
with Atomic Physics," Mason A. Porter and Predrag Cvitanović, Notices of the American
Mathematical Society, October, 2005.
w w w. a m s . o r g / m a t h m o m e n t s
MM/49
Boarding Faster
Waiting in line while boarding a plane isn’t just irritating, it’s also costly:The extra
time on the ground amounts to millions of dollars each year in lost revenue for
the airlines. Research into different boarding procedures uses mathematics such as
Lorentzian geometry and random matrix theory to demonstrate that open seating
is a quick way to board while back-to-front boarding is extremely slow. In
fact, mathematical models show that even people boarding at random get to
their assigned seats faster than when boarding back-to-front.
Figuring out your own strategy for boarding a plane is hard enough, but modeling
the general problem—which depends on many variables such as distance
between rows, amount of carry-on baggage, and passengers’ waistlines—is
substantially more complex. So researchers were pleased when they discovered
that their theoretical analysis confirmed simulations conducted by some airlines.
An added bonus to the research is that the mathematics used in the boarding
problem is similar to that used to improve a disk drive’s data input and output
requests. One clear difference: Data doesn’t try to carry on an extra bit.
w w w. a m s . o r g / m a t h m o m e n t s
MM/50
3OLVING #RIMES
4HE #"3
46 PROGRAM .5-"23 SHOWS MODERN MATHEMATICS AND MATHEMATICIANS
AT WORKBOTH INSTRUMENTAL EACH WEEK IN SOLVING AND PREVENTING CRIMES !LTHOUGH
THE SERIES IS FICTIONAL MANY OF THE SHOWS EPISODES ARE BASED ON TRUE STORIES )N
FACT STATISTICS COMBINATORICS AND GRAPH THEORY ARE JUST SOME OF THE MATHEMATICAL
FIELDS BEING USED TODAY BY REAL
LIFE INVESTIGATORS TO SOLVE ACTUAL CRIMES
/NE OF THE MOST IMPRESSIVE INSTANCES OF MATHEMATICS SOLVING A CRIME WAS A CASE
IN WHICH AN ALGORITHM PINPOINTED A SERIAL OFFENDERS LOCATION BASED ON THE SITES
OF PREVIOUS CRIMES 7HEN $.! SAMPLES CLEARED ALL THE SUSPECTS LIVING IN THE AREA
HOWEVER THE NATURAL CONCLUSION WAS THAT THE MATHEMATICS WAS UNSOUND 4HEN A TIP
LED INVESTIGATORS TO A DEPUTY WHO HAD BEEN ABOVE SUSPICION BECAUSE OF HIS JOB
AND WHO HAD LIVED IN THE TARGET AREA (E WAS EVENTUALLY ARRESTED AND SENTENCED
PROVING THAT CRIME DOESNT PAY BUT CHECKING YOUR ASSUMPTIONS DOES
&OR MORE INFORMATION h4HE (OUND OF THE $ATA 0OINTSv "RUCE 'RIERSON
0OPULAR 3CIENCE !PRIL
0HOTO BY 0ATRICK %CCLESINE #"3 ¥ #"3 "2/!$#!34).'