Development Of Bionic Artificial Neural Network Simulator

The purpose of this paper is to design and implement a bionic artificial neural network simulator,and this simulator is regarded as a simulation platform of algorithms realizing different bionic neural network,and analyse the correlation coefficient to achieve and verify the corresponding network functions ultimately.The This paper introduces the characteristics of bionic artificial neural network,as well as the realization of the algorithm and some basic network structure.Bionic artificial neural network is proposed on the basis of biology neural network self-growth,self-evolution,independent intelligence and a series of features.Its network structure refers to the infrastructure of biological neural network,and uses the algorithm to simulate the growth process of biological neural network.The whole process of growing and screening the network is achieved on a bionic artificial network simulator.The simulator provides an environment for the formation of the whole network,which can flexibly configure the function of neurons in the network,the type of neurons,the growth algorithm of the network,the size of the network space and the conditions of the network to stop the growth and the screening mechanism of the network.In addition,the simulator can realize the display of the three-dimensional interface in the whole network growth process.The display effect can help to locate the input and output neurons accurately,adjust the design of the network growth algorithm and modify the basic network structure.The size of the neurons during the whole network growth and evolution process is relatively large,and the calculation of each neuron is accelerated by the graphics processor.This is also the basis for the future expansion of the network size.This paper introduces the design of various functional modules of the bionic artificial neural network simulator and the functions that can be realized.The simulator provides an important tool for the study of bionic artificial neural networks for the growth of imitation flies and imitation bees.And the number of neurons in the whole bionic artificial neural network is flexibly configured by the different size of the neurons in the biological neural network such as drosophila and bee.The stimulus and growth mechanism of the whole network adopts the characteristics of spike neural network.The simulation of the spike neural network is very close to the biological neural network,and this network regards time which affects the network structure and performance as an important index also,which is the reason why spike neural network is an important reference for bionic artificial neural network.Designing of the bionic artificial neural network simulator make the bionic artificial neural network can be used to simulate or approximate the neural network structure of the drosophilas and bees.At the same time,according to the working principle of drosophilas' compound eyes and the basic network structure of bees,bionic artificial neural network can realize some network functions that identify or classify simple digital and pulse sequences.The results of these experiments have initially reached the expected target for the network function requirements.