COMPUTING WITH QUANTUMDOT CELLULAR AUTOMATA - Bio-Inspired and Nanoscale Integrated Computing - page 128

Normal input

Crossover input

Crossover output

Normal output

Figure 4.17. A multilayer crossover. Cells shown with an ''X'' are located on
higher layers; cells shown with ''O'' are vertically stacked to produce a via.

4.10. QCADesigner: SIMULATION ENGINES

Simulation engines within QCADesigner model the static and dynamic behavior
of QCA operation. Although two simulation engines are currently included with
the tool, other groups are free to incorporate their own simulation engines since
the source code for the tool is made available. Work on a new simulation engine
tailored for continuously clocked molecular QCA has been started [65]. However,
much work remains to be completed.
Models that described the static and dynamic behavior of electronic QCA
have been reported in [47, 51, 54, 67, 68]. In general, quantum mechanical systems
are not suitable for efficient simulation on a classical computer, and the models are
approximations of the full quantum mechanical behavior of the system. Included
in the current version of QCADesigner are two different simulation engines: the
bistable and the coherence vector engines. The bistable engine is based on the
simplest model and is very fast. There are several key assumptions made by this
model. The first is the two-state approximation, which ignores higher energy
configurations of the cell. The second is the Hartree approximation, which
quantum mechanically decouples the cells, ignoring the intercellular correlations
and reducing the size of the matrices involved. The last is that the cells remain at
the ground state, no time-dependant dynamics are considered. The second
simulation engine is based on the coherence vector formalism, a model that is
able to model the dynamics of the cells in the presence of coupling to the substrate.
This engine is also based on the two-state cell and uses the Hartree approximation
to decouple the correlations between cells.
The main challenge in implementing physically accurate simulations is the
lack of experimental data for the various QCA implementations. Even though
several small QCA systems have been developed as proof-of-concept experiments
[9, 10, 12, 20-22], these early devices do not necessarily represent scalable QCA