Evolvable Hardware
J.UCS Special Issue
Nadia Nedjah (Department of Electronics
Engineering and Telecommunications, Faculty of Engineering State,
University of Rio de Janeiro, Brazil)
nadia@eng.uerj.br
Luiza de Macedo Mourelle (Department of Systems
Engineering and Computation, Faculty of Engineering, State University
of Rio de Janeiro, Brazil)
ldmm@eng.uerj.br
Abstract: In recent years, genetic programming has attracted
many researcher's attention and so became a consolidated methodology
to automatically create new competitive computer programs. Concise and
efficient synthesis of a variety of systems has been generated by
evolutionary computations.
Evolvable hardware is a growing discipline. It allows one to evolve
creative and novel hardware architectures given the expected
input/output behavior. There are two kinds of evolvable hardware:
extrinsic and intrinsic. The former relies on a simulated evolutionary
process to evaluate the characteristics of the evolved designs while
the latter uses hardware itself to do so. Usually, reconfigurable
hardware such FPGA and FPAA are exploited.
While the intrinsic/extrinsic distinction is irrelevant for evolving
software, the differences in hardware are very apparent. With most
engineering hardware devices, evolution in simulation is quicker than
building many prototypes of a device, but electronic hardware can be
quicker to evolve intrinsically. This has the potential for producing
remarkably different results: whereas in simulation, a design is
naturally constrained to the programmer's model of the device,
intrinsic evolution allows a design to exploit natural features of the
device which may not even be understood by the programmer.
The concept of extrinsic was pioneered by Adrian Thompson who in 1996
evolved a tone discriminator, using fewer than 40 programmable logic
gates and no clock signal in a FPGA. This is a remarkably small design
for such a device, and it is still not understood how it works. For
example, one group of gates has no logical connection to the rest of
the circuit, yet is crucial to its function. It is thought that this
group somehow modulates the power supply or influences other
connections by generating a Magnetic field.
One of the main problems that still researchers are faced with in the
field of evolutionary hardware design is the scalability. Also, the
evolved circuits worked only on the original device and could not be
copied and they were only operational in a limited range of
temperatures.
These limitations constitute challenges for
researchers. Current research on evolvable hardware focuses on
discovering and applying some relevant biological concepts to get rid
of the so far imposed limitations.
This special issue of Journal of Universal Computer Science is devoted
to reporting innovative and significant progress in automatic hardware
design. It includes five contributed papers, whose main contributions
are described in the sequel.
In the first contribution, which is entitled "A Multi-objective Genetic
Approach to Mapping Problem on Network-on-Chip", Giuseppe
Ascia, Vincenzo Catania, and Maurizio Palesi address the problem of
topological mapping of intellectual properties (IPs) on the tiles of a
mesh-based NoC architecture. The aim is to obtain the Pareto mappings
that maximize performance and minimize power dissipation. The authors
propose a heuristic technique based on evolutionary computing to
obtain an optimal approximation of the Pareto-optimal front in an
efficient and accurate way. Besides, they extend two of the most
widely-known approaches to mapping in mesh-based NoC architectures in
order to explore the mapping space in a multicriteria mode. The
evaluation performed on both synthesized traffic and real applications
(an MPEG-4 codec) allowed the authors to confirm the efficiency,
accuracy and scalability of the proposed approach.
In the second contribution, which is entitled "Pareto-Optimal Hardware
for Substitution Boxes", Nadia Nedjah and Luiza de Macedo
Mourelle propose a methodology based on genetic programming to
automatically generate hardware designs for substitution boxes
necessary for many cryptosystems such as DES and AES encryption
systems. They aim at evolving minimal hardware specifications, which
minimize both space, response time and dissipated power. The author
claim that the evolved hardware compared with existing and well-known
designs, which were produced by human designers using conventional
methods, is superior.
In the third contribution, which is entitled "Automatic
Programming Methodologies for Electronic Hardware Fault
Monitoring", Ajith Abraham and Crina Grosan present and
apply three variants of Genetic Programming approaches to intelligent
online performance monitoring of electronic circuits and
systems. Obtained empirical results are compared with artificial
neural networks trained using back-propagation algorithm and
classification and regression trees. Based on this comparison, the
authors claim that the developed model which is GP-based could play an
important role for future fault monitoring systems.
In the fourth contribution, which is entitled "Multi-Objective
Evolutionary Algorithms and Pattern Search Methods for Circuit Design
Problems", Tonio Biondi, Angelo Ciccazzo, Santo D'Antona,
Giuseppe Nicosia and Salvatore Spinella experimentally compare the
effectiveness of multi-objective evolutionary algorithms and standard
optimization techniques applied to problems arising from Analog
Circuit Design: the optimization of a Operational Trans-conductance
Amplifier and of a fifth-order leapfrog filter. In particular, the
authors conclude that Pareto fronts determined by evolutionary
algorithms has a better spread of solutions with a larger number of
non-dominated solutions when compared to the classical multi-objective
techniques.
In the fifth contribution, which is entitled "DS/CDMA Multi-user
Detection with Evolutionary Algorithms", Fernando
Ciriaco, Taufik Abrão and Paul Jean E. Jeszensky analyse two
heuristic algorithms based on the genetic evolution theory applied to
direct sequence code division multiple access (DS/CDMA) communication
systems.
Based on
Monte Carlo simulation results, the authors guarantee that the
detection based on evolutionary heuristic algorithms is a viable
option when compared with the optimum solution (ML - maximum
likelihood), even for hostile channel conditions and severe system
operation.
As guest editors of this special issue of J.UCS, we wish to thank all
the contributors for their hard work and their promptness. Also, our
special thanks go to the reviewers that provided us with the right
feedback at the right time. This allowed us to fulfill our aim with
the expected quality and delays.
April 2006
Nadia Nedjah, Rio de Janeiro, Brazil
Luiza de Macedo Mourelle, Rio de Janeiro, Brazil
|