In the context of models of computation or of physical systems, quantum cellular automaton refers to the merger of elements of both the study of cellular automata in conventional computer science and the study of quantum information processing. In particular, the following are features of models of quantum cellular automata:
The computation is considered to come about by parallel operation of multiple computing devices, or cells. The cells are usually taken to be identical, finite-dimensional quantum systems.
Each cell has a neighborhood of other cells. Altogether these form a network of cells, which is usually taken to be regular.
The evolution of all of the cells has a number of physics-like symmetries. Locality is one: the next state of a cell depends only on its current state and that of its neighbours. Homogeneity is another: the evolution acts the same everywhere, and is independent of time.
The state space of the cells, and the operations performed on them, should be motivated by principles of quantum mechanics.
Another feature that is often considered important for a model of quantum cellular automata is that it should be universal for quantum computation. Models which have been proposed recently impose further conditions, e.g. that quantum cellular automata should be reversible and/or locally unitary, and have an easily determined global transition function from the rule for updating individual cells. Recent results show that these properties can be derived axiomatically, from the symmetries of the global evolution.
Models
Early proposals
In 1982, Richard Feynman suggested an initial approach to quantizing a model of cellular automata. In 1985, David Deutsch presented a formal development of the subject. Later, Gerhard Grössing and Anton Zeilinger introduced the term "quantum cellular automata" to refer to a model they defined in 1988, although their model had very little in common with the concepts developed by Deutsch and so has not been developed significantly as a model of computation.
Models of universal quantum computation
The first formal model of quantum cellular automata to be researched in depth was that introduced by John Watrous. This model was developed further by Wim van Dam, as well as Christoph Dürr, Huong LêThanh, and Miklos Santha, Jozef Gruska. and Pablo Arrighi. However it was later realised that this definition was too loose, in the sense that some instances of it allow superluminal signalling. A second wave of models includes those of Susanne Richter and Reinhard Werner, of Benjamin Schumacher and Reinhard Werner, of Carlos Pérez-Delgado and Donny Cheung, and of Pablo Arrighi, Vincent Nesme and Reinhard Werner. These are all closely related, and do not suffer any such locality issue. In the end one can say that they all agree to picture quantum cellular automata as just some large quantum circuit, infinitely repeating across time and space.
Models of physical systems
Models of quantum cellular automata have been proposed by David Meyer, Bruce Boghosian and Washington Taylor, and Peter Love and Bruce Boghosian as a means of simulating quantum lattice gases, motivated by the use of "classical" cellular automata to model classical physical phenomena such as gas dispersion. Criteria determining when a quantum cellular automaton can be described as quantum lattice gas automaton were given by Asif Shakeel and Peter Love.
A proposal for implementing classical cellular automata by systems designed with quantum dots has been proposed under the name "quantum cellular automata" by Doug Tougaw and Craig Lent, as a replacement for classical computation using CMOS technology. In order to better differentiate between this proposal and models of cellular automata which perform quantum computation, many authors working on this subject now refer to this as a quantum dot cellular automaton.
