I have been involved in a number of research projects throughout both my current Postdoctroal research at the Institute for Systems Research and my graduate studies at the University of Maryland at College Park as a member of the Center for Satellite and Hybrid Communication Networks (CSHCN). This gave me the opportunity to learn about a lot of different topics either individually or from my colleagues while working on the team projects. The following topics are the main subjects of research that I have been involved in. My advisor is Professor John S. Baras.

Broadcast Scheduling: In this work we focused on the problem of broadcast scheduling. This problem arises in the data delivery systems with broadcast capability, like satellite or wireless systems, when they intend to deliver the popular files such as web pages or information packages to a number of users simultaneously. In a system with several such files, the problem is that of determining the file that must be broadcasted at each decision instant in order to minimize the average waiting time of all customers. We formulated this problem as a Markov decision problem and found near-optimal policies for cases with identical file sizes and also when the files have different random or deterministic sizes. The latter result is particularly useful for cache broadcasting in satellite and wireless Internet systems. Currently, I am working on other variations of the problem. Minimizing the average waiting time over all users is mostly a requirement of the service provider. A stricter requirement imposed by the users is to associate each information page with a deadline and require the service provider to provide the data to the users within those deadlines. The problem I am currently working on is the broadcast scheduling in systems with deadlines where the objective is to minimize the total latency experienced by the users.

Wireless Ad-hoc Networks: One of the projects I am currently involved in is the developement of efficient algorithms for maintaining the connectivity across a MANET through deplyment of flying nodes. More specifically, as part of our DARPA-NMS project, we developed on-line algorithms for the placement of aerial platforms(e.g. UAVs) in order to establish the connectivity of mobile nodes. Our objective was to find the minimum number of required UAVs and find their movement trajectory based on the prediction of the movement of the MANET nodes. This problem is in fact an extended version of the Multiple Facility Location problem which is shown to be Non-polynomial in general. We formulated the problem as an optimization problem and presented an algorithm for finding a solution to it. The algorithm performs very close to optimal in different scenarios and I am currently working on extensions of the algorithm to provide more complicated QoS and connectivity requirements.

Media Access problems: The MAC layer protocols have been the subject of many research works for about three decades. However, there are still many unsolved problems in this field particularly when it comes to the interaction of MAC layer protocols with the protocols used in the other layers of the network. I am currently working on the performance evaluation of the 802.11 protocol where we calculate the delay and throughput of the user packets as a function of the number of users and also the type of the application generating the packets. Our analytical approach produces precise results that allows the system designer to predict the performance of the system before its implementation.

Traffic Modeling: The traffic in today's networks is undergoing rapid changes in terms of its composition. Therefore, there is a continuous need for constant monitoring of the behavior of the traffic at different levels and modifying the models that are assigned to its components, when necessary. I started this work with a detailed analysis of the network traffic at the hybrid gateway of the DirecPC system. I extracted various statistical properties of the traffic in different levels such as packet level and connection level. This study enabled us to determine the validity of different models for modeling the traffic in the network and in the queuing analysis of the system. The work on the traffic models was continued when I was awarded a fellowship from Lockheed Martin Communications and I worked with them in developing realistic traffic models for different applications like HTTP, FTP, Voice, MPEG Video, etc. for use in their simulations. I also worked with a group of students at CSHCN to develop proper simulation models for mobile ad-hoc networks. These models are used in various studies to evaluate the performances of different routing and scheduling algorithms in ad-hoc networks.