Design and application of Structural Formula Process Neural Network based on quantum evolutionary algorithm

Abstract

Aiming at the problems that the Structural Formula Process Neural Network (SFPNN) model has more study parameters, compute complexly after orthogonal basis expanding, and is difficult to converge. A quantum evolutionary algorithm is presented based on the quantum theory. The algorithm used the Pauli matrices to establish the axis of rotation, used qubits in Bloch sphere to rotate around the axis method to carry out optimal search, each particle represents three optimal solution to be updated at the same time, using the Hadamard gate achieve individual variability to avoid premature, enhancing the ergodicity of the solution space, expanding the search range of solution space, and approaching global optimal solution faster Taking network traffic and sunspot number prediction as an application, the simulation results show that the algorithm is validity.