AN INTRODUCTION TO NANOCOMPUTING - Bio-Inspired and Nanoscale Integrated Computing - page 18

In addition to being used for communication, waves can also be used for
computation by using interference. For example, constructive and destructive
interference can result in either a high intensity or low intensity—introducing the
digital abstraction without any switches. Furthermore, several waves can be
combined at a single point, an occurence that can be advantageous over
traditional logic that requires many transistors to handle many inputs. By
combining the benefits of waves over wires with the functionality of waves over
transistors, many limitations, theoretical and practical, can be overcome.
The visual portion of the electromagnetic spectrum has wavelengths on the
order of hundreds of nanometers, slightly larger than the nanoscale level. The
principles of diffraction and interference, however, apply to any type of wave,
including X-rays and electrons that have nanoscale or smaller wavelengths (recall
from Fig. 1.1). Another useful nanoscale wave, known as a spin wave, occurs when
the spin state of previous electrons affects the spin state of nearby electrons, causing
a propagation of the change in magnetic field (recall that magnetism is the
macroscopic property of spin). In addition to the many benefits of wave computing
described above, a key benefit of spin waves is that they can conveniently
communicate with electronic devices as well. Spin waves for computation are
described in Chapters 7, 8, 9, 14, and 19.
1.5.3. DNA and Protein Computing
DNA and protein are nanoscale molecular structures found in almost all existing
biological life that we know. Recall that DNA is essentially a sequence of four
primitive molecular structures: adenine (A), cytosine (C), guanine (G), and
thymine (T), attached to a molecular ''backbone.'' The main property that can
be exploited for computing is that adenine bonds only to thymine, and cytosine
bonds only to guanine. This means that a sequence of these base pairs has exactly
one complementary sequence that will bond to it. By setting up specific sequences
of DNA, many clever interlocking ''tiles'' can be created, and the way these tiles
interact is used for computation.
Proteins are complex molecules comprised of a string of amino acids.
Biologically, the sequence of amino acids that create a protein is defined by a
sequence of DNA. Various proteins can interact to perform useful computations,
for example, by exploiting the way a protein structure folds. Proteins are much
more complex than DNA, and their use for computation has so far only been
simulated [28].
The power of this paradigm is the principle of constraint satisfaction. Here,
instead of representing abstract operations as physical phenomena, abstract rules
are enforced (in this case with molecular structure), and physical phenomena (in
this case, bonding between structures) do their best to find the lowest state of
energy within the constraints. Performing constraint satisfaction in this way
makes it possible to compute many things that are otherwise very difficult. The
best example is a demonstration of DNA computing to solve the Hamiltonian
Path Problem [29]. This is a well known problem in theoretical computer science