On numerical simulations of integrate - and - re neural
نویسندگان
چکیده
It is shown that very small time steps are required to correctly reproduce the synchronization properties of large networks of integrate-and-re neurons when the diierential system describing their dynamics is integrated with the standard Euler or second order Runge-Kutta algorithms. The reason for that behavior is analyzed and a simple improvement of these algorithms is proposed.
منابع مشابه
Speed-up methods for simulations of traveling waves in integrate-and-fire neural networks
We consider single and multiple spike traveling waves solutions in an integrate-and-#re neural network. Due to the special properties of this system it is possible to deduce the dynamical state of each neuron in the full network from a small subset, using recurrence relationships. Since the dimension of the subset is on the order of the number of the traveling waves, this greatly reduces the co...
متن کاملPrediction of Entrance Length for Magnetohydrodynamics Channels Flow using Numerical simulation and Artificial Neural Network
This paper focuses on using the numerical finite volume method (FVM) and artificial neural network (ANN) in order to propose a correlation for computing the entrance length of laminar magnetohydrodynamics (MHD) channels flow. In the first step, for different values of the Reynolds (Re) and Hartmann (Ha) numbers (600<ReL increases.
متن کاملOn the convergence speed of artificial neural networks in the solving of linear systems
Artificial neural networks have the advantages such as learning, adaptation, fault-tolerance, parallelism and generalization. This paper is a scrutiny on the application of diverse learning methods in speed of convergence in neural networks. For this aim, first we introduce a perceptron method based on artificial neural networks which has been applied for solving a non-singula...
متن کاملPrediction of Pressure Drop of Al2O3-Water Nanofluid in Flat Tubes Using CFD and Artificial Neural Networks
In the present study, Computational Fluid Dynamics (CFD) techniques and Artificial Neural Networks (ANN) are used to predict the pressure drop value (Δp ) of Al2O3-water nanofluid in flat tubes. Δp is predicted taking into account five input variables: tube flattening (H), inlet volumetric flow rate (Qi ), wall heat flux (qnw ), nanoparticle volume fraction (Φ) and nanoparticle diameter (dp ...
متن کاملNumerical simulation of effect of non-spherical particle shape and bed size on hydrodynamics of packed beds
Fluid flow has a fundamental role in the performance of packed bed reactors. Some related issues, such as pressure drop, are strongly affected by porosity, so non-spherical particles are used in industry for enhancement or creation of the desired porosity. In this study, the effects of particle shape, size, and porosity of the bed on the hydrodynamics of packed beds are investigated with three ...
متن کامل