4 Simulation Analysis
In this section, we propose a large-scale complex networks containing 1,00,000 terminals, and all of them are freely connected to each other if necessary. Types of the
terminals are eNodeBs, IoT devices, signal processing cloud units, etc. Assume the
traffic model is Poisson, type of terminals is five, maximum connections for each nodes
are ten, and there is no cooperation and information exchange for data fusion. In Fig. 3,
we displayed the generated neural network in this scenario.
The model of the network is established using 3GPP Rel-13, the simulation platform is 3GPP LTE-R13 system-level simulation platform, and scenario is set to big
match. In Fig. 4, we provide the evaluation from four°: the connection, transmission,
Fig. 3. Generated neural network with five layers to discover and rank the importance of
parameters.
550
Y. Gao et al.
In this section, we propose a large-scale complex networks containing 1,00,000 terminals, and all of them are freely connected to each other if necessary. Types of the
terminals are eNodeBs, IoT devices, signal processing cloud units, etc. Assume the
traffic model is Poisson, type of terminals is five, maximum connections for each nodes
are ten, and there is no cooperation and information exchange for data fusion. In Fig. 3,
we displayed the generated neural network in this scenario.
The model of the network is established using 3GPP Rel-13, the simulation platform is 3GPP LTE-R13 system-level simulation platform, and scenario is set to big
match. In Fig. 4, we provide the evaluation from four°: the connection, transmission,
Fig. 3. Generated neural network with five layers to discover and rank the importance of
parameters.
550
Y. Gao et al.