Sajjad Mohsin

SDBATS: A Novel Algorithm for Task Scheduling in Heterogeneous Computing Systems

A heterogeneous computing system (HCS) efficiently utilizes the heterogeneity of diverse computational resources interconnected with high speed networks to execute a group of compute intensive tasks. These are typically represented by means of a directed acyclic graph (DAG) with varied computational requirements and constraints. The optimal scheduling of the given set of precedence-constrained tasks to available resources is a core concern in HCS and is known to be NP-complete problem. Task prioritization has been a major criterion for achieving high performance in HCS. This paper presents a SD-Based Algorithm for Task Scheduling (SDBATS) which uses the standard deviation of the expected execution time of a given task on the available resources in the heterogeneous computing environment as a key attribute for assigning task priority. This new approach takes into account the task heterogeneity and achieves a significant reduction in the overall execution time of a given application. The performance of the proposed algorithm has been extensively studied under a variety of conditions on standard task graphs from Graph Partition Archive as well as on some real world application DAGs such as Gaussian Elimination and Fast Fourier Transformation application DAGs. Our results show that SDBATS outperforms well known existing DAG scheduling algorithms in terms of schedule length (make span) and speedup.

Neural Network Based MIMO-OFDM Channel Equalizer Using Comb-Type Pilot Arrangement

MIMO (Multiple Input Multiple Output) and OFDM (Orthogonal frequency division multiplexing) bringing along a number of pros; a combination of both stands a good possibility of being the next-generation (4th generation) of mobile wireless systems. The technology however imposes a challenge that is the increased complexity of channel equalization. Wireless channels are multipath fading channels, causing deformation in the signal. To remove the effect (imposed by channel) from received signal, the receiver needs to have knowledge of CIR (Channel impulse response) that is usually provided by a separate channel estimator. This paper is aimed at exploring the use of Neural Network (NN) as a tool for MIMO-OFDM channel estimation and compensation. The research attempts to gauges the usefulness of proposed system by analyzing different algorithms to train NN. Further to ascertain the performance of the proposed technique; length of the known training sequence has been varied over a reasonable range and observations are made. Finally, the results obtained by using different algorithms for training NN have been compared with each-other and against the traditional least squares channel estimator, which along with observations/comments form part of the paper.

Optimization of Cylindrical Pin-Fin Heat Sinks Using Genetic Algorithms

In this paper, genetic algorithms (GAs) are applied for the optimization of pin-fin heat sinks. GAs are usually considered as a computational method to obtain optimal solution in a very large solution space. Entropy generation rate due to heat transfer and pressure drop across pin-fins is minimized by using GAs. Analytical/empirical correlations for heat transfer coefficients and friction factors are used in the optimization model, where the characteristic length is used as the diameter of the pin and reference velocity used in Reynolds number and pressure drop is based on the minimum free area available for the fluid flow. Both inline and staggered arrangements are studied and their relative performance is compared on the basis of equal overall volume of heat sinks. It is demonstrated that geometric parameters, material properties, and flow conditions can be simultaneously optimized using GA.

A Review on Demand Response: Pricing, Optimization, and Appliance Scheduling☆

Abstract The evolution of conventional electric grid into Smart Grid (SG) has enabled utilities as well as consumers to reap fruits due to its time varying price mechanisms. The utilities can acquire benefits by improving stability of grid, lessening blackouts and brownouts, knowing better their consumers power needs and not investing into new infrastructures. On the other hand consumer can also reduce electric bills, gain incentives by installing renewable energy sources and exporting energy to the main grid and attain improved services from utility. Demand Response (DR) is one of the most cost effective and reliable techniques used by utilities for consumers load shifting. In this paper, we are presenting a review of several DR techniques with a specific view on pricing signals, optimization, appliance scheduling used and their benefits. A comprehensive performance comparison is also prepared with the help of multiple criteria of SG paradigm.

Illumination normalization preprocessing for face recognition

This paper describes an illumination normalization technique which works at the pre-processing stage where the face image is first divided into equal sub-regions. Each sub-region is then processed separately for illumination normalization. Then the segments are joined back followed by further processing like noise removal and contrast enhancement. The proposed technique is tested on Yale dataset and compared with some previous illumination normalization techniques.

Urdu Spoken Digits Recognition Using Classified MFCC and Backpropgation Neural Network

Neural networks have found profound success in the area of pattern recognition. In the recent years there has been use of neural network for speech recognition. In this paper backpropagation neural network has been used for isolated spoken Urdu digits recognition. Mel frequency cepstral coefficients (MFCC) has been used to represent speech signal. Dimensions of speech features were reduced to a vector of 39 values. Only 39 values from MFCC features speech are fed to the neural network having more than one hidden layers with varying number of neurons, for training and recognition An analysis has been made between different number of hidden layers and different number of neurons on hidden layers. It has been found that results for these 39 values are similar to that obtained using complete MFCC features that range from 804 to 67x39. With the use of 39 values on input layer, computational complexity and time for training and recognition of neural network is reduced. In order to evaluate the significance of the proposed method on data other than Urdu digits, 30 English words have been trained and recognized that gave 98% results. All the implementation has been done inMATLAB.

Muhadith: A Cloud Based Distributed Expert System for Classification of Ahadith

This paper presents a novel approach for the classification of the religious scriptures, the Hadith (sayings of Prophet Muhammad (plural Ahadith)). Muhadith is a distributed, Cloud based expert system that uses the Hadith science to classify Ahadith among 24 types from seven broad categories. Classification of the Hadith is a complex and sensitive task, and can only be performed by an expert of the Hadith sciences. Muhadith expert system is designed to imitate the Hadith experts for Hadith classification, and to enable a computer to behave like a Hadith expert to discriminate the authentic Ahadith from unauthentic ones. This paper presents the relationship and mapping of the expert system technology onto Hadith sciences, and technicalities involved in designing of the Muhadith expert system. We also propose solutions for the communicational and interoperability problems faced by the legacy web based distributed expert systems. We employ service oriented architecture to overcome the communicational problem and a candidature for the Software as a Service (SaaS) for the Cloud computing. The expert system also provides a reasoning facility that enables the user to look into the classification details. Muhadith expert system has been designed by merging the ideas from the domains of expert systems, Web technologies, and distributed computing systems. This type of an effort on the topic is rare and applying them in the domain of Hadith is our humble contribution.

Adaptive image denoising using cuckoo algorithm

This paper presents a novel denoising approach based on smoothing linear and nonlinear filters combined with an optimization algorithm. The optimization algorithm used was cuckoo search algorithm and is employed to determine the optimal sequence of filters for each kind of noise. Noises that would be eliminated form images using the proposed approach including Gaussian, speckle, and salt and pepper noise. The denoising behaviour of nonlinear filters and wavelet shrinkage threshold methods have also been analysed and compared with the proposed approach. Results show the robustness of the proposed filter when compared with the state-of-the-art methods in terms of peak signal-to-noise ratio and image quality index. Furthermore, a comparative analysis is provided between the said optimization algorithm and the genetic algorithm.

Intelligent Image Retrieval Techniques: A Survey

In the current era of digital communication, the use of digital images has increased for expressing, sharing andinterpreting information. While working with digital images, quite often it is necessary to search for a specific image for aparticular situation based on the visual contents of the image. This task looks easy if you are dealing with tens of imagesbut it gets more difficult when the number of images goes from tens to hundreds and thousands, and the same contentbasedsearching task becomes extremely complex when the number of images is in the millions. To deal with thesituation, some intelligent way of content-based searching is required to fulfill the searching request with right visualcontents in a reasonable amount of time. There are some really smart techniques proposed by researchers for efficientand robust content-based image retrieval. In this research, the aim is to highlight the efforts of researchers whoconducted some brilliant work and to provide a proof of concept for intelligent content-based image retrieval techniques.

Efficient Way of Skull Stripping in MRI to Detect Brain Tumor by Applying Morphological Operations, after Detection of False Background

Brain MRI is used to get deeper view of the brain conditions. Skull stripping is a major phase sometimes refers to a pre-process in MRI brain imaging applications which refers to the removal of brain non-cerebral tissues. Various algorithms have been developed to improve the effectiveness of stripping skull from MRI. Morphological algorithms of “Erosion” and “Dilation” are recursively applied together to remove the skull. Besides the removal of skull, “erosion” distorts some cerebral tissues due to the presence of falsebackground. So “Dilation” process is applied for the restoration. In this study, we improved the efficiency of stripping skull in MRI using systematic application of “Erosion” with AOI (Area of Interest) approach after the detection of false-background. Before applying “Erosion”, a false back ground is detected. We identified the skull boundary through Dilation and then used scan line algorithm to fill the false background area. Consequently “Erosion” algorithm will only erode the AOI, resulting in the stripping of skull without any effect on the other tissues of the brain. Results show that the accuracy rate up to 95% is obtained and 43% efficiency is increased as compared to the different morphological techniques used previously.