Yadzhak M. Highly parallel algorithms and means for solving mass arithmetical and logical computations problems

The dissertation is dedicated to development of methods and highly parallel algorithms for solving of digital filtering problems and organisation of mass arithmetical and logical computations in multilevel neuron networks of feed forward and networks with projective and lateral links. The quasisystolic method for solving of digital filtering problems of any measurable is proposed. This method for solving of multiply cascade digital filtering problems is developed. The quasisystolic and synchronous methods for solving filtering problems are generalized for organisation of mass computations in neuron networks. The architecture of computing means to perform the proposed algorithms is developed.