Equation set solver based on memristor linear neural network and operation method of equation set solver

Abstract

The invention discloses an equation set solver based on a memristor linear neural network and an operation method of the equation set solver. The solver comprises a digital-to-analog conversion module, a memristor array, a current subtraction circuit, an analog-to-digital conversion module, a comparison module and a conductivity modulation module, the memristor serves as a neural synapse of the neural network, analog vector-matrix multiplication operation is executed, the training process of the memristor linear neural network is completed, and finally an equation set is solved. According to the invention, the matrix inverse of the coefficient matrix is solved in the training process of solving the neural network; when differential equation is solved, only one training process is needed; the numerical solution can be expanded to the whole time domain through iteration, the circuit complexity is reduced, the process of one data transmission is reduced, and the circuit consumption is reduced. Compared with the traditional process of solving a linear equation set by using a computer, the circuit has the advantages that the time complexity can be effectively reduced, the fusion of storage and calculation can be realized, the operation energy consumption and time can be greatly saved, and the reliability is high.

Description

technical field [0001] The invention belongs to the field of analog circuits, and more specifically relates to an equation set solver based on a memristive linear neural network and an operation method thereof. Background technique [0002] Linear equations are the basic mathematical models of many practical problems. At the same time, many differential equations that are difficult to obtain analytical solutions can be transformed into linear equations by finite difference method to obtain numerical solutions of differential equations, which further expands the application range of differential equations. , so an efficient and rapid linear equation solving system is needed to meet the needs of practical applications. [0003] The traditional system for solving linear equations is based on the von Neumann computer architecture. This computing architecture, which separates the memory from the calculator, frequently transmits data between the processor and the memory when solvi...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the defects of the prior art, the purpose of the present invention is to provide a memristive linear neural network-based equation solver and its operation method, which adopts the idea of ​​neural network to solve the delays existing in existing mathematical methods and computing circuits. The problem of high time and high power consumption

Images