Mohammed El-Beltagy

On the feasibility of centrally-coordinated Peer-to-Peer live streaming

In this paper we present an exploration of central coordination as a way of managing P2P live streaming overlays. The main point is to show the elements needed to construct a system with that approach. A key element in the feasibility of this approach is a near real-time optimization engine for peer selection. Peer organization in a way that enables high bandwidth utilization plus optimized peer selection based on multiple utility factors make it possible to achieve large source bandwidth savings and provide high quality of user experience. The benefits of our approach are also seen most when NAT constraints come into play.

Topologically evolved photonic crystals: Breaking the world record in band gap size

Using topology optimization, a photonic crystal (PtC) unit cell can be designed to exhibit favorable electromagnetic wave propagation properties. Among these is the opening of a band gap (BG) with the largest possible ratio of width to midgap frequency. In this paper the aim is to maximize the relative size of the first and fourth relative BGs of two-dimensional (2D) PtCs with a square lattice configuration. In addition, we examine the effects of the degree of unit cell symmetry on the relative BG size and on the geometric traits of the optimized topologies. We use a specialized genetic algorithm (GA) for our search. The results show that the type of symmetry constraint imposed has a significant, and rather subtle, effect on the unit cell topology and BG size of the emerging optimal designs. In pursuit of record values of BG size, we report two low-symmetry unit cells as an outcome of our search efforts to date: one with a relative BG size of 46% for TE waves and the other with a relative BG size of 47% for TM waves.

MyP2PWorld: Highly Reproducible Application-Level Emulation of P2P Systems

In this paper, we describe an application-level emulator for P2P systems with a special focus on high reproducibility. We achieve reproducibility by taking control over the scheduling of concurrent events from the operating system. We accomplish that for inter- and intra- peer concurrency. The development of the system was driven by the need to enhance the testing process of an already-developed industrial product. Therefore, we were constrained by the architecture of the overlying application. However, we managed to provide highly transparent emulation by wrapping standard/widely-used networking and concurrency APIs. The resulting environment has proven to be useful in a production environment. At this stage, it started to be general enough to be used in the testing process of applications other than the one it was created to test.

On the Optimality and Speed of the Deep Greedy Switching Algorithm for Linear Assignment Problems

The Deep Greedy Switching algorithm is a fast heuristic for solving large instances of the linear sum assignment problem whilst sacrificing very little in terms of optimality. In this paper we explore the worst case performance aspects of the algorithm. We prove that the algorithm is finite and analyze its computational complexity. We also discuss a number of simplified variations of the algorithm that shed some light on how the algorithm works. The basic algorithm is also extended to problems involving partial assignment. Computational results on a number of problem types show that the solutions reached by our algorithm are only slightly worse than those obtained by the auction algorithm. A parallelized version of the basic algorithm is presented. Our algorithm is significantly faster than other methods with approaching a comparable quality of solution.

Evolutionary Multiobjective Optimization for Portfolios in Emerging Markets: Contrasting Higher Moments and Median Models

Multi-objective Evolutionary algorithms are well suited to Portfolio Optimization and hence have been applied in complex situations were traditional mathematical programming falls short. Often they were used in portfolios scenario of classical Mean-Variance which are not applicable to the Emerging Markets. Emerging Markets are characterized by return distributions that have shown to exhibit significance departure from normality and are characterized by skewness and fat tails. Therefore higher moments models and median models have been suggested in the literature for asset allocation in this case. Three higher moment models namely the Mean-Variance-Skewness, Mean-Variance-Skewness-Kurtosis, Mean-Variance-Skewness-Kurtosis for return and liquidity and three median models namely the Median-Value at Risk, Median-Conditional Value at Risk and Median-Mean Absolute Deviation are formulated as a multi-objective problem and solved using a multi-objective evolutionary algorithm namely the non-dominated sorting genetic algorithm II. The six models are compared and tested on real financial data of the Egyptian Index EGX. The median models were found in general to outperform the higher moments models. The performance of the median models was found to be better as the out-sample time increases.

Project Portfolio Exploration and Visualization using Self-Organizing Maps

When dealing with portfolio analysis and assessment, decision makers often have to select a portfolio of a subset of proposed projects with varying levels of investment. The process involves comparing large sets of project portfolios across many dimensions. A tool that would make such comparisons cognizable would greatly facilitate portfolio selection. We tackle this challenge by visualizing the set of candidate portfolios using a self-organizing map (SOM). We adopt and modify the matching and updating steps of the SOM to work with project portfolio data. Our approach is based on content aware distance calculation between the data and codebook vectors as well as a novel process of updating the codebooks when dealing with project portfolio data.

Optimal Design of Periodic Timoshenko Beams using Genetic Algorithms

A periodic material is composed of two or more types of elastic materials laid out in space in a repeated periodic fashion. The properties of a periodic composite material depend on the arrangement of two or more materials within its unit cell. Periodic materials (also known as phononic materials) are commonly characterized by its dispersive frequency spectrum. With appropriate spatial distribution of the constituent material phases, spectral stop bands could be generated. Moreover, it is possible to control the number, the width, and the location of these bands within a frequency range of interest. This study aims at exploring the relationship between the unit cell conguration and its frequency spectrum characteristics. Focusing on 1D layered phononic materials (in the form of rods and beams), and longitudinal wave propagation in the direction normal to the layering, the unit cell features of interest are the number of layers and the material phase and relative thickness of each layer. An evolutionary search for multi-phase cell designs exhibiting a wide stop band, or a series of wide stop bands, is conducted using a specially formulated representation and set of operators that break the symmetries in the problem. Structures composed of the designed phononic materials are excellent candidates for use in a wide range of applications including the development of vibration and shock isolation structures, sound isolation pads/partitions, multiple band frequency lters, among many other applications.

NAT Constraints Management in Tree-Based P2P Live Streaming Systems

In this paper we present an approach for organizing peers in a centralized tree-based P2P live streaming system in order to minimize peer connectivity failures due to network address translation (NAT) constraints. The centralized coordinator in our system optimizes peers position according to their upload bandwidth to reach maximum P2P contribution and optimizes the position of each single peer to avoid incompatible NAT connectivity between peers which can cause a decrease in the systems performance. A network optimization technique is used to reach the optimal peer placement in the tree to satisfy the download demands of as many peers as possible.

Exploring altruism in social networks

We show in this paper how altruistic behaviour emerges in social networks from an entirely selfish population. The analysis is conducted by simulating the ultimatum game on a Barabasi-Albert network with different degrees of preferential attachment. In previous work pair wise random matching was used in selecting players where all agents are connected, while in the proposed model interactions are limited to only connected agents on the graph. The analysis shows how a social networks structure limits the prevalence of the altruistic behaviour compared to previous work, and how a decrease in selection frequency will increase the chances of selfish agents to survive.

Design Space Exploration of Multi-Phase Layered Phononic Materials via Natural Evolution

A phononic material is commonly characterized by its dispersive frequency spectrum. With appropriate spatial distribution of the constituent material phases, spectral stop bands could be generated. Moreover, it is possible to control the number, the width, and the location of these bands within a frequency range of interest. This study aims at exploring the relationship between the unit cell configuration and its frequency spectrum characteristics. Focusing on 1D layered phononic materials, and longitudinal wave propagation in the direction normal to the layering, the unit cell features of interest are the number of layers and the material phase and relative thickness of each layer. An evolutionary search for multi-phase cell designs exhibiting a wide stop band, or a series of wide stop bands, is conducted using a specially formulated representation and set of operators that break the symmetries in the problem. An array of optimal designs for a range of ratios in Youngs modulus and density are obtained and the corresponding objective values are plotted as a function of the ratios of the phase properties. Structures composed of the designed phononic materials are excellent candidates for use in a wide range of applications including vibration and sound isolation.Copyright