2012
WAMS 2012 Proceeding
Feed forward neural network and simulation of filter adaptation by DAPHNIA
Authors: Tibor Kmet, Maria Kmetova
Abstract
A model of feeding adaptation of a filter feeder is presented. Based on the assumption that filtration adaptability represents optimization type process is incorporated. Two possible strategies were followed an instantaneous optimality at each time interval and an integral formulation, maximization of the integral biomass which leads to the optimal control problem with control and state constraints. The optimal control problem is transcribed into nonlinear programming problem, which is implemented with adaptive critic feed forward neural network and recurrent neural network for solving nonlinear projection equations. Also stability analysis of equilibria and some numerical simulation is given. It is shown that Hopf bifurcation may occur depending on filtration rate.
I3M Scientific Sponsors
I3M Industrial Sponsors
I3M Media Sponsors
