International Journal of Hybrid Intelligence (10 papers in press)
Intentional and Unintentional Misbehaving Node Detection and Prevention in Mobile Ad hoc Network
by Arshad Ahmad Khan
Abstract: MANET (Mobile Ad-hoc Network) is a flexible and autonomous network, which is composed of heterogeneous mobile nodes disseminated in a wireless communication area, in which nodes move freely and are arranged in a random approach. Mobile nodes produce the dynamic and uncertain network topology. Adaptation and self-forming are the two main qualities of MANET that create various types of sensitive and critical applications in various fields such as healthcare, vehicular communication, and military, These fields are highly sensitive and a single incorrect message may directly or indirectly affect human life. Hence, data communication in this area should be secure. One of the major resources for data communication is routing. Routing protocol establishes Route selection in MANET based on the assumption that all the nodes present in a network are cooperative, but this assumption is not valid in a hostile environment. Misbehaving activities in network layer are caused by packet drop or refusal to forward the packets by intermediate nodes. However, packets may also be dropped by intermediate nodes due to constrained resources. In this paper, initially we study the routing misbehavior activities in MANET. Further we propose the mechanism to determine and mitigate the malicious node accurately by avoiding all possible reasons for packet drop due to constrained resources. Analytical and simulation results show that the proposed work mitigates misbehaving node with less overhead in comparison with existing secure knowledge algorithm
Keywords: MANETs; misbehaving node; constraint resources; routing.
Collaboration adaptive filtering model for data reduction in Wireless Sensor Networks
by Walaa M. Elsayed, Hazem M. El-Bakry, Salah M. El-Sayed
Abstract: Wireless sensor networks (WSNs) are collecting data periodically by randomly dispersed sensors (motes) that typically exploit high energy in monitoring a specified application. Furthermore, dissemination mode in WSN is resulting noisy or missing information that affects the behaviour of WSN. So, data prediction-based filtering is an important approach to reduce redundant data transmissions, conserve node energy and overcome the defects resulted from data dissemination. Therefore, in this article, a Distributed Data Reduction Model (DDRM) is proposed to prolong the network lifetime by decreasing the energy consumption of sensor nodes. It is built upon a distributive clustering model for predicting diffusion-faults in WSN. The proposed model is developed using the RLS adaptive filter integrated with a FIR filter for minimizing the amount of transmitted data and provide high convergence of the signals. A dataset of atmospheric changes was handled. The results clarify that DDRM reduced the rate of data transmission to ~ 20 %. Also, it depressed the energy consumption to ~ 95 % throughout the dataset sample. DDRM effectively upgraded the performance of the sensory network by about 19.5 %, and hence extend its lifetime.
Keywords: WSN; Cluster head; Data dissemination; Adaptive RLS filter; Data prediction; Value failure.
Multi Area Power Dispatch Strategy Considering Economic and Environmental Aspects Using NDSGA II
by Abhik Hazra, Saborni Das, Ashish Laddha, Mousumi Basu
Abstract: Multi area economic environmental dispatch strategies (MAEEDS), corresponding to centralized fossil fuel fired power plants, remain significant for allocating power among committed units of various regions in an optimalized manner. The allocation must ensure simultaneous reduction in total fuel cost and emission level in the best possible ways. At the same time, equality alongside inequality constraints likes production-demand balance, power production capacity, and tie line capacity have to be taken into consideration. The aforementioned optimalization task involves multiple objective optimization with contradictory behaving goals. The presented article suggests nondominated sorting genetic algorithm II (NDSGA II) to achieve solutions for the MAEEDS task. Solutions obtained through the NDSGA II for a four area test system prove the superiority of the suggested technique. A comparison has also been made among the solutions attained through the suggested technique, strength pareto evolutionary algorithm II (STPEA II), and other well-known optimization techniques available in the literature. The comparative analysis shows superiority of the suggested NDSGA II for the considered four area MAEEDS task.
Keywords: Multi area economic environmental dispatch strategies; multiple objective optimization; nondominated sorting genetic algorithm II; strength pareto evolutionary algorithm II.
Special Issue on: Hybrid Computational Intelligence in Big Data Analytics and Cloud Computing
Smart Healthcare Model with Fog-Cloud Network Architecture
by Rohan Basu Roy, Arani Roy, Amitava Mukherjee, Mrinal Kanti Naskar
Abstract: In this paper, the performance of an Internet of Things (IoT) based, real time, smart ECG signal compression and transmission protocol is investigated through both simulation and hardware implementation. The protocol consists of a combination of Fog and Cloud computing architecture. The input to the model is noisy ECG signal acquired through sensors. The model is based upon a four layered structurernwith first layer consisting of wearable ECG sensor and noise filter embedded in the device. Second layer is an encoder which consists of an algorithm subdivided into two schemes namely, Geometry Based Method (GBM) and Wavelet Based Iterative Thresholding (WTIT). The algorithm is based on the fact that ECG signals can be approximated by the linear combination of a few coefficients obtained from a wavelet basis. GBM reduces the minimal signal values to zero geometrically in time domain and WTIT encodes the signal in time-frequency domain. Further, Compressed Row Hoffman Coding algorithm (CRHFC) is applied to convert the sparse coefficient matrices to compressed, transmittable matrices. The third layer consists of wireless transmission medium from the wearable to a private Cloud, from where data is accessed by the hospital servers. The data obtained in the receiver is the final layer where signal reconstruction is performed using inverse transforms. The performance metrics-compression ratio (CR), percentage RMS differencern(PRD), Quality Score (QS), time complexity and sparsity are used to evaluate the performance of the model.
Keywords: IoT; hardware implementation; Fog; Cloud; compression.
Improved real coded genetic algorithm based short-term hydrothermal generation planning
by DIPANWITA GANGULY, Saborni Das, Abhik Hazra, Ashish Laddha, Mousumi Basu
Abstract: Real coded genetic algorithm (RCGA) and improved real coded genetic algorithm (IRCGA) have been applied for the solution of short-term hydrothermal scheduling problem. The improved technique has been developed and tested on a multi reservoir cascaded hydroelectric system having generation-load power balance, upper and lower limits on reservoir capacity, water discharge rate, water spillage rate, hydraulic continuity restriction and operating capacity limits of different hydro and thermal units. The water transport delay between connected reservoirs has also been taken into consideration. The performance of the proposed approach is validated with four test systems. The results of the proposed algorithm are compared with those of modified differential evolution (MDE), teaching learning based optimization (TLBO), clonal selection algorithm (CSA), improved fast evolutionary programming (IFEP), improved particle swarm optimization (IPSO), and genetic algorithm (GA). From numerical results, it has been found that the IRCGA based approach is able to provide better solutions in lesser computational time.
Keywords: Hydrothermal operation planning; improved real coded genetic algorithm; limiting values of ramping rate; loading effect of valve point; restricted operating sections.
A Survey on Big Data: An Emerging Imparity and Revolution in Digital World
by Anupam Mukherjee, Sourav De, Siddhartha Bhattacharyya
Abstract: Size of the heterogeneous data is increasing rapidly at an electrifying speed. But we cannot handle this massive amount of unstructured data in traditional database, most of the data in the digital universe is unstructured. Big data analytics provides better computational power and efficient mechanism to handle this situation. This paper attempts to offer a survey report of big data, which changes rapidly of high volume, velocity, verity. This survey paper considers some of the major challenges of big data and its characteristics followed by a conceptual framework of big data. This survey paper also focuses on the complexity of medical images, data mining and soft computing problems.
Keywords: Big data analysis; Social media data; unstructured data; heterogeneous dimensionalities; architectural framework; medical image data; data mining; soft computing.
A comparative study of text mining in big data analytics using Deep Learning and other machine learning algorithms
by Souvik Chowdhury, Shibakali Gupta
Abstract: Text mining has become an important aspect in modern day. Text mining has various applications e.g. spam email classification, similar news item clustering etc. There have been many scenarios where regression is accompanied with text mining e.g. for predicting sales of a product of any store the product description also plays an important role. Big data analytics applications modify information scientists, prognostic modelers, statisticians and different analytics professionals to investigate growing volumes of structured group action information, and different sorts of information that square measure typically left untapped by standard business intelligence (BI) and analytics programs. That encompasses a mixture of semi-structured associate degreed unstructured info -- as an example, web clickstream info, net server logs, social media content, text from shopper emails and survey responses, mobile-phone call-detail records and machine info captured by sensors connected to the net of things. In this paper we have tried to solve text mining problems in big data analytics using Deep Learning methods. Deep learning on the other hand known to be strongest supervised learning method. Here we can make use of back propagation concept to harness the power. We can also use gradient descent learning method to reduce the cost function and settle to global minima. We will also do a comparative analysis of other machine learning algorithms with deep learning methods. We will also construct a rXs contingency table popularly known as Crosstab.
Keywords: Deep Learning; Text mining; Machine Learning.
Cloud Database Failure Prediction using Multi Agent System
by Souvik Chowdhury, Shibakali Gupta
Abstract: We have proposed a new method of fault tolerance mechanism in modern cloud database. Cloud database is important for big enterprises and have strict SLA (Service Level Agreement) so any downtime is costly. All databases have their own fault tolerance mechanism by means of clustering technique, standby, backup recovery strategies with almost no data loss. The above methods are either costly or time consuming. If any measurement can be taken which can predict cloud database failure well ahead of it actually happens, suggest reason for the failure and also provide solution for the problem from Knowledge database will not only reduce cost and time problem mentioned above but also will add a new dimension to fault tolerance technology. In this paper we have tried to develop an idea of failure prediction for an Cloud database with the help of Multi Agent System and probable solution based on Knowledge base search.
Keywords: Cloud database; Oracle; Multi Agent System (MAS); RDBMS.
IoT based algorithms for distributed location detection for flights
by Amlan Chatterjee, Hugo Flores, Soumya Sen, Khondker Hasan, Ashish Mani
Abstract: Detecting the location of aircraft at all times during flight is of utmost importance in commercial aviation. In recent times, there have been instances of aircraft that became untraceable during flight, and has not been located. This potentially has a huge negative impact on the entire aviation industry, and recovery of specific airlines from such an issue is extremely difficult. With the increasing number of aircraft that operate over various routes across the globe, tracking each flight with accurate location detection is a challenge. Specifically, over oceanic routes, where aircraft are not within the range of radar or other traditional tracking devices, the issue of location tracking is difficult with existing infrastructure. In this paper, we propose an internet of things based framework for aircraft that can assist with tracking the same. Our introduced model includes aircraft, air traffic control towers, project loon balloons and infrastructure enabled aircraft as the different components. Our proposed algorithms work for aircraft that operate on routes over land as well as large water bodies, specifically oceans. The algorithms are distributed in nature, thereby creating a model with no single point of failure. All the introduced algorithms are implemented and tested using simulation for the tracking of aircraft in different scenarios. The experimental results show that using the proposed techniques, an additional 70% of aircraft under consideration can share location data and can be detected as compared to using conventional radar based techniques.
Keywords: Internet of Things; IoT; Distributed Location Detection; Commercial Aviation; Communication; Flight Tracking.
Combined Economic Emission and Load Dispatch Using Hybrid Metaheuristics
by Dipankar Santra, Krishna Sarker, Anirban Mukherjee, Subrata Mondal
Abstract: This paper attempts to report optimal result for economic emission dispatch (EED) problem together with economic load dispatch (ELD) problem with valve point effect. The result is obtained using a hybrid meta-heuristic technique not used earlier for the said purpose. The technique involves hybridized particle swarm optimization (PSO) and ant colony optimization (ACO) methods. The purpose is to minimize fuel cost and operating cost of generators and at the same time minimize harmful emission of NOX. In this study ELD problem has been considered for a power system with 13 generating units whereas ELD and EED both combined i.e. CEED problem is solved for 40-generator test system. The proposed PSO-ACO hybrid shows excellent convergence property and encouraging results in comparison with other hybrid methods.
Keywords: combined economic emission dispatch (CEED); economic load dispatch (ELD); fossil fuel; environment; hybrid metaheuristics; PSO; ACO.