IIT Bhubaneswar Student Registration:: All accepted papers will be published online in the proceeding of SCOPUS indexed AISC series of Springer :: Some of the extended/modified selected quality papers will be published in a Special Issue of 'Swarm and Evolutionary Computation journal, Elsevier (SCI)'

Abstracts

Abstract :

The Shannon/Nyquist sampling theorem tells us that in order to not lose information when uniformly sampling a signal we must sample at least two times faster than its bandwidth. In many applications, including digital image and video cameras, the Nyquist rate can be so high that we end up with too many samples and the data must be compressed for storage or transmission. In many applications, including imaging systems (medical scanners, radars) and high-speed analog-to-digital converters, increasing the sampling rate or density beyond the current state-of-the-art is very expensive. In this lecture, we will mention a new technique that tackles these issues using compressive sensing... We will replace the conventional sampling and reconstruction operations with a more general linear measurement scheme coupled with an optimization in order to acquire certain kinds of signals at a rate significantly below Nyquist rate. We will mention the links between data acquisition, linear algebra, inverse problems, compression, dimensionality reduction, and optimization in a variety of situations where knowledge based systems have been found to be very useful. We will show results of a case study for surveillance.

By:

R. N. Mohapatra
Mathematics Department
University of Central Florida
Orlando, Florida, 32817, USA

Abstract :

Rapid advance of various technologies associated with unmanned aerial vehicles (UAVs) has enabled many complex tasks to be carried out autonomously with no human intervention. UAVs can be deployed for numerous applications such as reconnaissance and surveillance, traffic monitoring, detection and containment of hazardous leakages in industries and so on. Both hovering capability and agility of quad-copter UAVs (popularly known as “drones”) comes very handy in many of these applications. However, for harnessing the true potential, it is quite obvious that drones should have a good built-in mechanism and associated guidance logic to autonomously land successfully and on the designated target with high precision.

This talk will essentially present a bio-inspired Tau-guidance approach for smooth and precision landing of drones. The approach is essentially based on the Tau theory, which has been established by studying the landing behaviour of birds, thereby generating a desired reference trajectory. Next, the differential geometric (dynamic inversion) philosophy is used to generate the necessary guidance commands to the drone. Experimental studies were performed using commercially available low-cost AR-Drone. To overcome the navigational error associated with the sensors, built-in monocular SLAM (Simultaneous Localization And Mapping) and INS (Inertial Navigation System) data were fused using a Kalman Filter, which improved the navigation accuracy remarkably. To improve the accuracy even further in the final terminal area, the landing site was supported by a range-sensing Kinect camera. With this low cost navigation solution, hardware experimental results show that autonomous landing of drones is possible with high precision using the bio-inspired Tau-theory. Details of the Tau-guidance philosophy and the results obtained will be presented in this keynote.

By:

Radhakant Padhi
Professor, Dept. of Aerospace Engineering
Indian Institute of Science, Bangalore

Abstract :

For over past one decade or so, several "so called" novel metaheuristic optimization algorithms are being proposed on a regular basis, by using nature both as model and metaphor. However, whether this horde of algorithms will make any real impact to the progress of non-convex optimization or whether their influence will remain stipulated within the world of paper writing - this question can be answered only through careful benchmarking and testing of these algorithms. This talk will address the modern benchmarking procedures, advanced performance indices and most importantly the statistical hypothesis testing procedures usually applied for judging the significance of the comparative studies that we most often find in evolutionary computing literature. The talk will also discuss some very important future research issues that need attention from the practitioners of swarm and evolutionary metaheuristics. The discussions of the talk will remain confined to population-based metaheuristics for continuous search spaces and for single-objective bound-constrained optimization.

By:

Swagatam Das
Professor, Electronics and Communication Sciences
Indian Institute of Science, Bangalore

Abstract :

Pricing financial derivatives, as an important and modern branch of applied mathematics and an interdisciplinary area between finance and mathematics, has posed many challenging soft-computing issues to mathematicians and computer scientists. For instance, cleverly designing computational algorithm for option pricing, developing analytical approximations to avoid unnecessarily waste of computational resources and calibrating models effectively and efficiently are several important areas of recent interest as far as soft computing is concerned.

This talk consists of three parts. In the first part, a comprehensive review is given on various quantitative approaches for pricing financial derivatives, particularly American options, under various models with pros and cons of each approach being briefly discussed. In the second part, some analytical approximations are particularly focused to demonstrate soft-computing in financial derivative pricing. Finally, numerical solution approaches adopted in pricing American options are discussed with some most recent research in terms of model calibration presented here as well.

By:

Song-Ping ZHU
School of Mathematics and Applied Statistics,
University of Wollongong,NSW 2522, Australia

Abstract :

Most practical problem solving tasks in engineering, sciences and business involve various complexities that are difficult to handle using simplified classical methods. These methods make a modification to the original problem in order to suit their application. Thus, a solution by these methods may not be appropriate to the practical task. Nature's way of arriving at robust and adaptable organisms to survive and thrive in various complex environmental scenarios has been utilized to construct efficient optimization methodologies for solving various practical problems, as they are. In this talk, we shall present single and multi-objective evolutionary methods and demonstrate their usefulness on a number of practical problems.

By:

Kalyanmoy Deb
Koenig Endowed Chair Professor,
Michigan State University, East Lansing, USA