International Journal of Research and Reviews in Applied Sciences
ISSN: 2076-734X, EISSN: 2076-7366

Volume 29, Issue 1 (October, 2016)

To read and print the PDF files of the Journal Archive you will need to have Acrobat Reader 
 If you have any technical or content problems contact : publisher@arpapress.com

1. RELIABILITY OF REPAIRABLE k–out–of–n: F SYSTEM HAVING DISCRETE REPAIR AND FAILURE TIMES DISTRIBUTIONS
by Sevcan Demir Atalay & Özge Elmastaş Gültekin
Abstract

In this paper, repairable k – out – of – n: F system with single repairman is studied. The failure and repair times of components are assumed as geometric distribution which is one of the discrete distributions. Both the generalized and steady state transition probabilities of the states for the related system are obtained. The recursive relations for the reliability of the system are generated. Also, using geometric transformation technique, the MTSF (Mean Time System Failure) of the system is obtained. As an example of this system, 3– out – of –5: F system is considered. The transition probabilities, reliability and MTSF of this system are obtained and some graphs for the reliability and MTSF are drawn for different values of failure and repair time parameters.

Source: International Journal of Research and Reviews in Applied Sciences 
October 2016-- Vol. 29 Issue 1 -- 2016

2. BACKPROPAGATION NEURAL NETWORK APPROACH FOR MEAN TEMPERATURE PREDICTION

by Manal A. Ashour, Somia A. ElZahaby & Mahmoud I. Abdalla

Abstract

Temperature is one of the basic components of the weather. In this paper, mean temperature have been forecasted using Artificial Neural Network (ANN). The design of the ANN based on four weather parameters. The ANN design has been applied for Cairo city, the capital of Egypt. The training and testing used meteorological data for twenty years (1996- 2016). In this study we predict the mean temperature by using the artificial neural network ANN model and the multiple linear regression MLR model. This study provides a neural network model based on backpropagation to predict the mean temperature and to compare the obtained results with the results obtained by the multiple linear regression MLR model. The different performance evaluation criteria are introduced to compare the results obtained by the neural network model and the results obtained by the multiple linear regression. Results show improved performance of the neural network model over the multiple linear regression model.

Source: International Journal of Research and Reviews in Applied Sciences 
October 2016-- Vol. 29 Issue 1 -- 2016

3. TRAINING FEED FORWARD NEURAL NETWORK USING GENETIC ALGORITHM TO PREDICT MEAN TEMPERATURE
by Manal A. Ashour, Somia A. AlZahby & Mahmoud I. Abdalla
Abstract The genetic algorithm has been used to train fixed neural networks to predict the mean temperature. In this study the implementation of a genetic algorithm to neural network for the weather parameters prediction is considered. Various genetic operators to design a genetic algorithm are examined, with an objective of minimizing the cumulative square error. The effect of the crossover and mutation operators and population size are taken into account. By computer’s simulations, the effectiveness of different crossover and mutation operators in training the network are shown. The appropriate crossover operators are described with the different selection methods which are used. A design of experiment (DOE) approach is used to find the value of the genetic algorithm parameters through determining the probability of the crossover This study discusses three cases. The first case is when the roulette wheel selection RWS method and power mutation with SPOX, TPOX, AMOX, HOX, BLX, and LX crossover operators are applied one at a time. The second case is when the tournament selection TOR method and power mutation with the previously mentioned crossover operators are applied one at a time. And finally the third case is when the rank based fitness BRF method and power mutation with the previously mentioned crossover operators one at a time. Results show improved performance of TPOX+TOR+POWRE model over other models.
Source: International Journal of Research and Reviews in Applied Sciences
October 2016-- Vol. 29 Issue 1 -- 2016

4. THE STUDY OF CLUSTERING ALGORITHM BASED ON DENSITY DISTRIBUTION
by Shaohua-Zhou, Haipeng-Gu & Xiaoping-Ren
Abstract

The article is based on DBSCAN (Density-Based Spatial Clustering of Applications with Noise). Firstlythrough the adjustment of the parameters radius Eps and density threshold Minpts, which makes objects with high density in the data space formed an initial cluster, while the left objects are classified into a cluster waiting to be evaluatedThen according to the proposed clustering method based on gravity-based clusteringselecting K optimal clusters from the initial cluster so that all initial clusters can conduct the second clustering in order to form K initial clusters. Finally, according to the judgment, the evaluated objects waiting to be classified to the nearest initial class, the clusters come to an end. This algorithm decreases cluster results’ over-dependency on parameters; it can adapt different shapes of data and greatly improve the accuracy of clustering.

Source: International Journal of Research and Reviews in Applied Sciences
October 2016-- Vol. 29 Issue 1 -- 2016

5. EMERGY ANALYSIS OF THE PERI LAKE SYSTEM AND THE ROLE OF THE NEOTROPICAL OTTER
by Oldemar Carvalho Junior
Abstract

The main objective of this work is to evaluate the natural capital and ecosystem services of the Peri Lake System and the role of the Neotropical otter (Lontra longicaudis) as a flag species to change adverse realities, based on system ecology and emergy synthesis. First, a system diagram is constructed to organize the thinking and the relationships between components and pathways of exchange and resource flow. It is an overview of the system, combining different sources of information and organizing the efforts. The second step is to construct the emergy synthesis tables of flows directly from the diagrams. It accounts the annual flows of material, energy, and information that support the system, such as materials, energy, and information. Finally, emergy indices are calculated in order to summarize and relate emergy flows of the economy with those of the environment. Quantities of stored emergy of environmental resources are calculated from the sum of the emergy of all inputs and then multiply them by the time it takes to accumulate the storage. To calculate the emergy of economic storages, all inputs of energy, materials, and labor to produce them will be summed. The objective is to be able to predict the economic and environmental viability. For the evaluation of renewable inputs to the Peri Lake, solar energy, rainfall, runoff, and wind annual averages will be used. Transformities and specific emergies are calculated for biodiversity and endangered species. They are calculated by first quantifying all the emergy used in making the product or service and dividing by the energy of the product or service. The units can be in sej/J if the product is divided by the energy or sej/g if the emergy of the product is divided by the mass. Emdollar is a measure of the money circulating in the economy as a result of the emergy flow. The emdollar is obtained by dividing the total emergy driving the economy by the economy's Gross Domestic Product (GDP). Main services with market values provided by the Peri Lake are research, water supply, and fish harvest. The main economic activities are related to water supply, fishing, and tourism.

Source: International Journal of Research and Reviews in Applied Sciences
October 2016-- Vol. 29 Issue 1 -- 2016