Convergence Speed Of Network Research Articles

A New Levenberg Marquardt based Back Propagation Algorithm Trained with Cuckoo Search

Back propagation training algorithm is widely used techniques in artificial neural network and is also very popular optimization task in finding an optimal weight sets during the training process. However, traditional back propagation algorithms have some drawbacks such as getting stuck in local minimum and slow speed of convergence. This research proposed an improved Levenberg Marquardt (LM) based back propagation (BP) trained with Cuckoo search algorithm for fast and improved convergence speed of the hybrid neural networks learning method. The performance of the proposed algorithm is compared with Artificial Bee Colony (ABC) and the other hybridized procedure of its kind. The simulation outcomes show that the proposed algorithm performed better than other algorithm used in this study in term of convergence speed and rate.

Adaptive Synchronization of Complex Dynamical Networks with State Predictor

This paper addresses the adaptive synchronization of complex dynamical networks with nonlinear dynamics. Based on the Lyapunov method, it is shown that the network can synchronize to the synchronous state by introducing local adaptive strategy to the coupling strengths. Moreover, it is also proved that the convergence speed of complex dynamical networks can be increased via designing a state predictor. Finally, some numerical simulations are worked out to illustrate the analytical results.

Transformer Fault Diagnosis of Rough-Neural Network Based on MEA

Combining mind evolutionary algorithm with rough set and neural network, this paper proposed a rough set neural network based on MEA for transformer fault diagnosis. Rough set attribute reduction as the front-processor of neural network diagnostic device, and using MEA to search rough set discrete breakpoints and optimize neural network weights and thresholds, it avoided complex manual trial of the conventional rough set attribute reduction and slow convergence speed and low precision of BP neural network, then faster convergence to the global optimum solution and improves the diagnosis speed and accuracy. Simulation results show that this method is effective.

The Application on the Forecast of Plant Disease Based on an Improved BP Neural Network

Aiming at the disadvantages of large computing, slow convergence and easily trapping into local minima of traditional BP network, a new method named batch momentum learning algorithm which combining the momentum with batch gradient descent algorithm has been used to be as the learning algorithm of connection weights and threshold of BP neural network, through using this method to forecast the prevalence of plant disease, the convergence speed of BP neural network has been enhanced.

The Forecast of Logistics Cost of Coal Enterprises Based On PSO-BP

In order to forecast the logistics cost of coal enterprises effectively,the PSO-BP and BP network are respectively used to forecast the logistics cost ,the result shows that the PSO-BP network's convergence speed and prediction accuracy are obvious better than BP network. The PSO-BP neural network is very attractive for a wide application in forecasting logistics cost in coal enterprises.

OLSR performance measurement in a military mobile ad hoc network

Wireless ad hoc networks are autonomous, self-configurating and adaptive. Thus, such networks are excellent candidates for military tactical networks, where their ability to be operational rapidly and without any centralized entity is essential. As radio coverage is usually limited, multihop routing is often needed; this is achieved by an ad hoc routing protocol supporting nodes mobility. In this paper, we present performance measurements of the Optimized Link State Routing (OLSR) routing protocol, having the status of IETF RFC. The measurements are performed at CELAR site on a platform representative of military scenarios in urban areas. This platform consists of ten routers, eight PDAs and laptops using a IEEE 802.11b radio interface and implementing OLSR v7. Some nodes are mobile within vehicles. The emphasis of the measurements is on the performance of the network (route repair, network convergence speed, user traffic performance) in presence of this mobility.

Performance of a neural network method with set partitioning

Computational experience with a neural network method for the set partitioning problem (i.e. minimize cx subject to Ax = 1, x j = 0 or 1, where A is a zero-one matrix, 1 is a column of ones, and c is a non-negative integral row) is presented. The approach maps the variables (subsets) of the problem onto equal numbers of neurons of a neural network model of Boltzmann machine-type operation. By mapping a neuron onto each subset rather than on each variable x ij = 0 or 1 depending on whether element i is in subset j or not (and then use the method of ‘penalization of constraints’, as was used by Hopfield for the travelling salesman problem [J.J. Hopfield, Proc. Natn. Acad. Sci. USA 81, 3088–3092 (1984); J.J. Hopfield and D.W. Tank, Biol. Cyber. 52, 141–152 (1985)]), the complexity of the neural network is considerably reduced. We propose an efficient encoding formalism which determines a good global behaviour of the model, around the valid partial or feasible solutions of the implemented problem. An analysis of the network's energy yields several interesting results that allow us to anticipate good performance. Simulations have been performed on sets of artificial data and real data encountered in the airline crew scheduling real application with very large sizes and very low density. Our numerical study, which extends to 8000 subsets, exhibits an impressive level of solution quality when the involved parameters are chosen with care. By exploring the central problem of scaling behaviour with respect to the size of the instances handled by our method, we gain some insight into the relationship between solutions quality, optimum values of the network tuning parameters, convergence speed and instances size.

An artificial neural network-based codebook search technique

An unsupervised learning artificial neural network (ANN) [T. Kohonen, Helsinki University of Technology Report TKK-F-A601 (1986)] was modified for a vector-quantized PCM (VPCM) codebook search problem. The ANN was structured such that each element of the network had associated with it a single codebook vector. The amount of processing required at each element of the network was used to derive an upper bound on the number of network iterations. Simulations were performed to determine the effect that various network parameters had on the speed of network convergence. The parameter values offering the greatest performance were applied to the network. The speed and computational complexity of the ANN solution to this problem were then compared to the same criteria for a standard linear (full-codebook) search technique. Analysis and test results indicate that the ANN approach can provide the speed of a tree search coupled with the minimum memory characteristics associated with a linear search, at the expense of requiring a multiprocessor configuration. [Work supported by NSF.]