A design method of a linear equation solver based on a variable parameter convergent neural network

• Main Authors: ZHANG ZHIJUN, FU ZHENG, ZHENG LUNAN
• Format: Patent
• Language: Chinese; English
• Published: 18.12.2018
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention discloses a design method of a linear equation solver based on a variable parameter convergent neural network, comprising the following steps: 1) establishing a mathematical model of a practical physical system or a numerical solution system in the form of a smooth time-varying linear matrix equation in a real number domain; 2) obtaining a time-varying parameter matrix of the mathematical model through the sensor of the system in the step 1), and solving the time derivative thereof through a differentiator; 3) designing an error function equation of the system; 4) converging theneural network method with variable parameters in the real number domain and the obtained time-varying parameter matrix and derivative thereof, using the monotonically increasing singular excitation function to design a real domain smooth time-varying linear matrix equation solver and obtain the unique optimal solution of the real domain smooth time-varying linear matrix equation of the system. The execution end of the sy