Jan'19
Welcome to the IUP Journal of Computer Sciences
Focus
Social media might one day offer a dazzling, and even overwhelming, array of source material for historians. Such an abundance presents a logistical challenge (the total number of tweets ever written is nearing half a trillion) as well as an ethical one (will people get to opt out of having ephemeral thoughts entered into the historical record?). But this plethora of new media and materials may function as a totally new type of archive: a multidimensional ledger of events that academics, scholars, researchers and the general public can parse to generate a more prismatic recollection of history. [...]
How an Archive of the Internet Could Change History, 2016
Machine learning algorithms have created the sensation of working with
experts in an endless list of fields�advertisers, detectives, doctors, chess
masters, soccer players, loan creditors, salespeople, entertainment critics, drivers and even matchmakers. Machines with intelligence have dazzled the human mind with what they can do with data, and how much better they are at it. These smart algorithms can learn from our observations about the world and accomplish tasks that would appear to be insurmountable hurdles to humans. However, machine learning algorithms may be terrible in situations when the conditions are not right for them, for example, when they face an adversary. Adversarial learning often becomes an arms race between the learner and the adversary as the competition continues.
The first paper, �Optimized Link Stability and Cross-Layer Design for Secured Ad-Hoc Network,� by Anita Sethi, Sandip Vijay and J P Saini, discusses different cross-layer architecture and analyzes different parameters for improving the performance of the network.
The next paper, �A Comparison of Some Metaheuristic Optimization Algorithms for Solving High Dimensional Benchmark Problems�, by Parimal Kumar Giri and Chandrakant Mallick, presents three metaheuristic algorithms which are thoroughly analyzed and tested in detail for high dimensional real-parameter optimization problems. The algorithms are tested on three standard optimization benchmark functions.
Michael Gallagher, Erick Rengifo and Rossen Trendafilov have presented a four-part paper. This issue carries the first part of the paper, �Cloud Computing for MATLAB and R Users: Part I � Creation of Virtual Computing Instance Using Amazon Web Services�. Part I describes how to configure the virtual machine step-by-step and demonstrates how to use publicly available preconfigured virtual machines. Part II � Transferring Data and Cloud Storage, Part III � Multicore Virtual Machine Parallel Computing Environment, and Part IV � How to Use MATLAB in the Cloud of the paper will appear in the subsequent issues of this journal.
The short note, �A Proposal for Interactive Programming Language,� by Lincoln Hannah, presents an idea for a programming language specifically designed for an interactive environment like a notebook.
Consulting Editor
| Article | Price (₹) | ||
| Optimized Link Stability and Cross-Layer Design for Secured Ad-Hoc Network |
100
|
||
| A Comparison of Some Metaheuristic Optimization Algorithms for Solving High Dimensional Benchmark Problems |
100
|
||
| Cloud Computing for MATLAB and R Users: Part I � Creation of Virtual Computing Instance Using Amazon Web Services |
100
|
||
| A Proposal for Interactive Programming Language |
100
|
||
Optimized Link Stability and Cross-Layer Design for Secured Ad-Hoc Network
Maintainability and context awareness knowledge increase the requirement of cross-layer architecture. The wealth of cross-layer coordination examples from miscellaneous fields of networking are security, handoff, autonomic communication, routing, sensor networks, quality of service and energy. In the cognitive radio, selection of suitable wireless channels in the vicinity depends on application requirements and network conditions. Different static and dynamic cross-layer optimization heuristics are available in literature with different performance parameters. The paper discusses different cross-layer architecture and link optimization for improving the performance of network. The prime focus is on link prediction methodology for enhancement of the performance of network. Various performance parameters of protocols of software defined networks are compared. Throughput, end-to-end delay, jitter, PDR and goodput with different scenarios are observed and compared. It is observed that the performance of OLSR is better in different scenarios as compared to other protocols.
A Comparison of Some Metaheuristic Optimization Algorithms for Solving High Dimensional Benchmark Problems
With the emergence of Big Data, the existing optimization techniques need to be tested to find those that are not suitable to handle high dimensional problems. The number of natureinspired population-based metaheuristic optimization algorithms has been explored over the decade with new techniques being proposed constantly. A recent summary of existing algorithms has listed near about 134. A majority of these optimization algorithms have been designed and applied to solve real-parameter function optimization problems, each claimed to be superior to other methods in terms of performance. However, most of these algorithms have been tested on relatively low dimensional problems, i.e., problems involving less than 30 parameters. This paper presents different benchmark functions that are systematically analyzed and tested in detail for problems involving up to 100 parameters. Genetic Algorithms (GA), Biogeography-Based Optimization (BBO) and Particle Swarm Optimization (PSO) are compared in terms of accuracy and runtime using three high dimensional standard benchmark functions.
Cloud Computing for MATLAB and R Users: Part I � Creation of Virtual Computing Instance Using Amazon Web Services
Cloud-based computing has tremendous potential for academia in general and researchers in particular. The ability to store, read and manipulate large dataset is increasingly important to be on the cutting-edge of research. While most researchers have extensive programming skills, they generally are not computer scientists. A number of cloud-based and virtual work spaces have been developed which can bring enormous computing advantages over any desktop or laptop. However, the instructions to get up and running in the cloud are often overwhelming, and require a significant investment of time. There is an initial steep learning curve associated with developing a high performance computing environment that several times becomes an obstacle that prevents many to benefit the most from using the cloud. This paper seeks to demystify the process and get the reader up and running very quickly. Once this initial process is over and the reader is proficient on these issues, performing computational chores that would take days on a desktop can be done in minutes. We next show the way the readers can implement their research using two well-known computer programs: MATLAB and R.
A Proposal for Interactive Programming Language
The paper presents an idea for a programming language specifically designed for an interactive environment similar to a notebook. It aims at combining the power of a programming language with the usability of a spreadsheet. The comments presented could be modified as per needs.