Accuracy vs simulation speed trade-off enhancements in the generation of chaotic attractors

In this paper, the trade-off between accuracy and simulation speed in the generation of muti-scroll chaotic attractors at 1-D is analyzed and improved. In a first step, a macromodel based on bipolar transistors and passive elements is used to model the behavior of Opamps, and later on they are used to approach the behavior of all nonlinear system. Hspice simulations are executed to generate chaotic attractors in the phase plane and time-series. CPU-time used during the solution of the chaotic system is computed. In a second stage, a simple and accurate nonlinear model, which includes the most influential performance parameters for Opamps, is coded in C++ and it is used to generate a nonlinear system of equations, which models the behavior of all chaotic system. CPU-time is also computed. Because chaotic waveforms generated have a random behavior, statistical tests are used to measure the similitude/accuracy between two random variables during a long time. Our results indicate that chaotic waveforms can swiftly be generated by using the simple but accurate model for Opamps without the accuracy worsens, independently of the initial conditions of the chaotic system.