V.A. Astapenko, P.A. Golovinsky
In the modern communication networks the electromagnetic fields are used. These fields are strong enough for the description in terms of classical physics. The technical equipment of transmission and processing of the information by classical electromagnetic fields is based on the determined devices which opportunities are gradually exhausting.
One of the basic features of a new quantum signal is its probabilistic nature. Quantum physical state, transmitted by such signal, can be measured with a probability. Quantum states possess fundamental property of interference. Quantum calculations are probabilistic and have fundamental analog nature, in contrast with classical discrete calculations. An alternative to the coherence for all the process of calculations is its coherence during small space-time intervals.
For realization of partial coherence, the wave packages of Rydberg atomic states are considered. Effective way of creation of such states is excitation of atoms by a short laser pulse. The further control can be carried out by an electric, magnetic or high-frequency electromagnetic field. The interference of Rydberg wave packages provides a new procedure of comparison of the data. Coherent properties of wave functions of quantum systems are applicable for parallel computing processes connected to recognition of the complex data by new way. Advantage of quantum networks in comparison with usual networks of incoherent signals is the opportunity of simultaneous operating with huge information containing in amplitudes and relative phases of the quantum state components. In the paper, the pure states are considered. The mixed states can be used for combination of the coherent computing procedures and Markovian processes in future.
The basic difference of the developed approach from the algorithms of the usual quantum computers constructed on operations with qubits, consists in its analog nature and reduction of requirements to coherence. The important feature of the considered quantum networks is an opportunity of physical realization of separate steps of work inside small space-time intervals. It is feasible on a basis of already available technical solutions. The most perspective element, specified for the quantum neural networks, is based on quantum dots.