Computer Engineering

This study introduces a inflated multi-layered clustering protocol for ad hoc wireless sensor networks (WSNs), where the size of clusters is variable. So that the closer clusters to the base station (BS) have a smaller size than farther ones. Moreover, in each cluster, using some intelligent fuzzy rules and in a decentralized way, a novel sub tree strategy is determined. In this way, some parent nodes are chosen that are responsible for collecting and aggregating data from their adjacent or dinary nodes and sending them to its cluster head, directly or via other parent nodes, which substantially decreases intra-cluster communication energy costs. Furthermore, these two compatible techniques can fairly mitigate the hot spot problem resulting from multi-hop communication with the BS. The simulation results demonstrate that the proposed protocol outperforms two energy-efficient protocols named DSBCA and LEACH in terms of functional network longevity for both small-scale and large-scale sensor networks.

Unified Modeling Language has been widely used in software development for modeling the problem domain to solution domain. The major problems lie in comprehension and testing which can be found in whole process. Program slicing is an important approach to analyze, understand, test and maintain the program. It is a technique for analyzing program by focusing on statements which have dependence relation with slicing criterion. Program slicing is of two types (i) Static slicing (ii) Dynamic slicing. Dynamic slicing refers to a collection of program execution and may significantly reduce the size of the program slice because runtime information, collected during execution, is used to compute the program slice. In this paper we introduce an approach for constructing dynamic slice of unified modeling language (UML) using sequence diagram, state chart diagram, class diagram along with the activity diagram. First we construct an intermediate representation known as model dependency graph. MDG combines information extracted from various sequence diagrams along with those from class, state-machine, sequence diagram and activity diagrams. Then dynamic slice is computed by integrating the activity models into the MDG. Finally for a given slicing criterion test cases has been generated.

In this age of technology, smartphones play a vital role in almost all fields of social life to make it easy going a convenience. Day by day, the users of smartphones are increasing. There is no conditional boundation for using the smart phone. People, who own these devices tend to use them at their maximum as these devices such as mobile phones, are very convenient to use anytime, anywhere. Single application can use multi-platform means convenient for everyone. This paper tries to convey information about the current and earlier news events for the frequent users of Hadas-Eritrea and Eritrean profile, the end-users can be able to interact with more graphical features of this multi-platform mobile application. In this paper, we have proved Broadcasting Mobile Multi-Platform Hadas – Eritrea and Eritrean profile (BHMMP) with different platforms such as Windows7 OS, Android, iOS, and Windows Phone using PhoneGap/Apache Cordova framework.

As the demand increases for day-day applications, rapid transfer of high amount of data over high speed networks must be required also Bandwidth must be high enough for these applications. Congestion Control is an important subject relevant to these applications to maintain stability for any kind of network. In this paper, review on various congestion control mechanisms and their performance measurement parameters are to be compare with each other. Active Queue Management is one of the method to get control over congestion by dropping packets from buffer queue as an indication to other end node to slow down transfer of packets.

The health care field contain large amount of data and in order to process this data, we must use any advanced techniques that will be helpful to deliver effective results and make effective decisions on the data and obtain relevant results. Heart disease is a major problem and is one of the main reasons for saying no. of deaths occurring worldwide. In this paper, the practical framework of Heart Disease Prediction is applied using algorithms in Machine Learning such as Logistic regression, Naïve Bayes, Support vector machine, KNN, decision tree, random forest, XG-Boost and the neural network. This framework uses 13 factors such as age, gender, blood pressure, cholesterol, oldpeak, cp, etc. In the first step, we upload a database file and select an algorithm to perform on the selected database. Then accuracy is predicted for each selected algorithm and graph, and the model is designed for the one with the highest accuracy by training the database in it. In the next stage, input is given to each candidate parameter and based on that method produced, the stage with heart disease is predicted. We then take precautionary measures by looking at the patient's condition. Our strategy is effective in predicting the heart attack of a traumatized person. The Heart Disease Prediction Framework developed in this concept is one of the different methods that can be used within the heart disease category.

A new hybrid approach of Genetic Algorithm (GA) with local search algorithm is proposed to solve Course Timetabling Problems (CTP). In which, GA architecture is enhanced by proposing various selection, crossover and mutation operators. Diversity in population helps to get global optimum. In order to accommodate diversity of population and to avoid local optima, grade selection and combinatorial partially matched crossover operators are proposed. To increase the convergence rate and to produce guaranteed result, various mutation strategies are proposed with gene tuning approach. To improve the quality of the solution, steepest ascent hill climbing local search algorithm has been proposed. With these, hybrid approach with enhanced GA is implemented on CTP and hence its quality is proved by getting more promising and consistent results in all operations of the possible twelve combination of GA proposed operators. Also, proved experimentally that combination of grade selection, combinatorial partially matched crossover and adaptive mutation strategy operators is performing the best among all twelve proposed combinations and a combination of operators from the literature by yielding the average relative convergence rate as 31% which is greater than all others’ convergence rate.

Software Defined Radio (SDR) is a radio technique that has the capability of replacing several receivers with a single universal receiver. It includes a Digital Front-End (DFE) with the ultimate goal to implement all processing in digital domain. As the receiver has to adapt to various communication standards with different characteristics, the objective is to develop an optimum detection algorithm to combat Multiple Access Interference (MAI). Subtractive Interference Cancellation (IC) detectors like SIC and PIC are proposed and are employed in both uplink and downlink transmissions. Suitability of linear detectors like MRC and MMSE based MUD is being analyzed in multistage receiver.

Due to the development of new techniques for generating and collecting data, the rate of growth of scientific databases has become large, which creates both a need and an opportunity to extract implicit knowledge to analyze these datasets. Analysis of such large expression data gives rise to a number of new computational challenges not only due to the increase in no. of data objects but also due to the increase in no of attributes. Hence to improve the efficiency and accuracy of mining task on high dimensional data, the data must be preprocessed by an efficient dimensionality reduction method. In this paper, we have proposed a Rough Genetic Approach for high-dimensional data clustering. Initially an efficient method of Rough Set Theory has been applied on the discritized data set to generate a reduced set of relevant attributes. Then, it is proposed to use the Genetic Algorithm for finding the cluster index of the dataset with reduced attribute which may give better clustering accuracy than other clustering techniques.

The power consumption of modern microprocessors is a primary design limitation across all processing domains, from embedded devices to high performance chips. Reducing feature sizes and increasing numbers of transistors residing on a single die only impairs this issue. Schemes are immediately required to handle this issue, yet still deliver high performance from the system. Architectures have several options for the organization cache. Typically, there are two approaches to improve the efficiency of the cache: either by increasing the number of “live” blocks residing in the cache —while keeping the capacity of the cache intact— or by dynamically adjusting the cache size to the actual requirements of the executing applications. In this paper we have compared some proposed methodologies from selective research papers for improvement of cache performance. The first one resizes cache and unused part is gradually decayed i.e. stop producing hits. In second approach cache is divided into hierarchy, which is searched in sections one by one to get a required hit. Finally, in third approach we re-configure our cache by using a configuration vector that is loaded with a new configuration before an application is started.

In this paper a novel face identification technique, ACOG algorithm has been proposed which is a hybrid of ACO (Ant Colony Optimization) algorithm and GA (Genetic algorithm). The ACO processes and extracts the features of the input image over which several pre-processing steps are done to enhance the chances of feature extraction. The extracted features are given as input to GA which detects the face features and compares the features with the existing face database.