A New Method of Embedded Fourth Order with Four Stages to Study Raster CNN Simulation

A new Runge-Kutta (PK) fourth order with four stages embedded method with error control is presented in this paper for raster simulation in cellular neural network (CNN) environment. Through versatile algorithm, single layer/raster CNN array is implemented by incorporating the proposed technique. Simulation results have been obtained, and comparison has also been carried out to show the efficiency of the proposed numerical integration algorithm. The analytic expressions for local truncation error and global truncation error are derived. It is seen that the RK-embedded root mean square outperforms the RK-embedded Heronian mean and RK-embedded harmonic mean.