Search

The seismic design of new structures in Iran is currently based on earthquake hazard maps for a fixed return period of 475 years. However, this can lead to an unequal level of risk due to uncertainty in collapse capacity and differences in hazard curves in different locations. In the work reported in this paper, a risk-targeted approach was used to determine the peak ground acceleration (PGA) for designing reinforced concrete moment-resisting-frame (RC MRF) buildings in four important port cities in southern Iran. The three main parameters considered for determining risk-targeted PGA were the structural fragility curve, the seismic hazard curve and the target annual collapse probability. Nine RC MRF buildings with different geometrical characteristics were evaluated and their fragility curves were extracted. Then, using a risk integral, the annual collapse probability and PGA of the frames were obtained for the studied cities. Comparison of the results showed that an approach proposed in the literature for Europe is in accordance with the results of the approach proposed in this paper. [ABSTRACT FROM AUTHOR]

In this paper, a novel constructive-optimizer neural network (CONN) is proposed for the traveling salesman problem (TSP). CONN uses a feedback structure similar to Hopfield-type neural networks and a competitive training algorithm similar to the Kohonen-type self-organizing maps (K-SOMs). Consequently, CONN is composed of a constructive part, which grows the tour and an optimizer part to optimize it. In the training algorithm, an initial tour is created first and introduced to CONN. Then, it is trained in the constructive phase for adding a number of cities to the tour. Next, the training algorithm switches to the optimizer phase for optimizing the current tour by displacing the tour cities. After convergence in this phase, the training algorithm switches to the constructive phase anew and is continued until all cities are added to the tour. Furthermore, we investigate a relationship between the number of TSP cities and the number of cities to be added in each constructive phase. CONN was tested on nine sets of benchmark TSPs from TSPLIB to demonstrate its performance and efficiency. It performed better than several typical Neural networks (NNs), including KNIES_TSP_Local, KNIES_TSP_Global, Budinich's SOM, Co-Adaptive Net, and multivalued Hopfield network as wall as computationally comparable variants of the simulated annealing algorithm, in terms of both CPU time and accuracy. Furthermore, CONN converged considerably faster than expanding SOM and evolved integrated SOM and generated shorter tours compared to KNIES_DECOMPOSE. Although CONN is not yet comparable in terms of accuracy with some sophisticated computationally intensive algorithms, it converges significantly faster than they do. Generally speaking, CONN provides the best compromise between CPU time and accuracy among currently reported NNs for TSP. Index Terms--Constructive-optimizer neural network (CONN), Hopfield-type neural networks (HNNs), Kohonen-type self-organizing maps (K-SOMs), traveling salesman problem (TSP).

In this paper, a new shell clustering method is presented to cluster model-based shells in 2-dimensions. Shells, that each one of them can be expressed by a center and non-negative radius in each angle (by considering polar coordinate system), can cluster by using the proposed model-based fuzzy c-shells (MFCS) clustering method. In this paper, firstly one of the most famous traditionally clustering methods, i.e. fuzzy c-spherical shells (FCSS) clustering method, is extracted from the proposed MFCS clustering method as a specific state of it. Then, the performance of proposed method is examined in three examples when it is applied over shells with various shapes in 2-dimensions. Since the resulted systems of equations in the studied examples cannot be solved directly, the particle swarm optimization (PSO) algorithm is used to numerically solve the resulted equations systems. The simulation results show the acceptable performance of the proposed MFCS method. [ABSTRACT FROM AUTHOR]

In this paper a novel approach for cardiac arrhythmias detection is proposed. The proposed method is based on using independent component analysis (ICA) and wavelet transform to extract important features. Using the extracted features different machine learning classification schemas, MLP and RBF neural networks and K-nearest neighbor, are used to classify 274 instance signals from the MIT-BIH database. Simulations show that multilayer neural networks with Levenberg-Marquardt (LM) back propagation algorithm provide the optimal learning system. We were able to obtain 98.5% accuracy, which is an improvement in comparison with the similar works.