Dr. C.S. Lui, John - Selection Scheme:

  1. Selection Criteria : genuine interests in my projects, dedication, hard-working, willing to explore and learn new things.
  2. Time Slot for Interview: I have not set up any time slot. I believe in the open-door policy that any students can come in to ask question. Since I am in the office most of the time, you can stop by my office to inquire about the project, then we can talk about the final year projects in detail.

LUI9801

Title: Personal Organizer and Information/Resource Allocator in the Internet
Supervisor: Dr. John C.S. Lui
Number of students: 2
Project Type: CS or CE
Prerequisite: love system building, like to learn mathematics like probability and Bayesian Network
Project description: This project is about information searching, gathering, classification and exploration.

Imagine the following scenario. John Lui is a mad scientist. When he comes into his office, he likes to read electronic newspaper. Specifically, in areas like world news, education news, economic news and sport news. Therefore, he likes his personal information system (PIS) to automatically retrieve and organize the news before he comes into the office. John Lui also likes to schedule his daily and weekly appointments and meetings and he wants his PIS to remind him about his commitment. Since John is working on scheduling algorithms for O.S. system, he likes the PIS to automatically search for any related information in the Internet on a daily basis (therefore, the PIS needs to search, classify and categorize the information). Once a week, John likes to receive a video clip from the University president, who mulitcasts his video speech to all lecturers in CUHK once a week. Therefore, the PIS needs to allocate the resource (e.g. storage, and transmission bandwidth) for John so that he can view the video with high quality. Last but not least, based on John's profile, the PIS systems needs to inform John about some possible information that is of interest to John.

To do this project, the students need to learn about Bayesian Network (a form of probability reasoning), Java and Internet programming, Internet protocols and be very motivated and independent.

References:
  1. Personal Information Service (PIS)--An Application of Wide-Band Communications, 2012 A.D. Victor O.K. Li, April 1998, Volume 86, Number 04, IEEE Proceeding.
  2. For more information, talk to John Lui.

LUI9802

Title: Multicast Service for Low Earth Orbit (LEO) System
Supervisor: Dr. John C.S. Lui
Number of students: 2
Project Type: CS or CE
Prerequisite: Love to learn, love to think, need to read some related papers and device and implement the algorithms. Note that networking knowledge is not necessary.
Project description: One way to provide a global wireless communication is via the low earth orbit (LEO) satellite system. There are several proposed and on going projects, with major companies like Motorola and Microsoft supporting this expensive infrastructure. Since a LEO has a dynamic and time-varying connectivity pattern, therefore, it is important to reconsider the routing protocols so as to satisfy the reliability and quality of service constraints. The problems complicates further if we want to use the LEO system to provide multicast service (e.g. conferencing and meeting) to different mobile users. Therefore, source and participants all have time varying locations.

This final year project is to first read some papers related to the system ( I will study and read the papers with the students), and then we will device algorithms to support such type of multicast service.

References:
  1. Probabilistic Routing Protocol for Low Earth Orbit Satellite, Huseyin Uzunalioglu, Georgia Tech, USA.
  2. A Connection Handover Protocol for LEO Satellite ATM Networks, Huseyin Uzunalioglu, W. Yen, I.F. Akyildiz, Georgia Tech, USA.
  3. For more information, talk to John Lui.

LUI9803

Title: Parallel Simulation Algorithms: A mathematical and system perspective
Supervisor: Dr. John C.S. Lui
Number of students: 2
Project Type: CS or CE
Prerequisite: love mathematics, love to learn new thing, motivated and independent.
Project description: In order to quantify any newly proposed protocol, new architecture or new algorithm, we need to evaluate the performance gain of these new proposals. In general, there are several approaches to evaluate the proposed system, either via analytical methods or by simulation.

Simulation is considered as a very general tool (e.g. any person who knows how to program can do it). However, one of the major short coming is that it usually takes a long time for the execution of the simulation program. To overcome this, many research proposed to use parallel simulation method.

The aim of this project is to study and learn (together with me ) some parallel simulation methods. By the end of this project, we will build some "medium-sized" simulations so as to verify the parallel simulations methods. This project is ideal for some students who have interests in mathematics.

References:
  1. A Comparative Study of Parallel Algorithms for Simulating Continuous Time Markov Chains'', ACM Transactions on Modeling and Computer Simulation, vol. 5, no. 4, October 1995, pp. 326-354.
  2. Optimistic Parallel Simulation of Markov Chains Using Uniformization, Journal of Parallel and Distributed Computing, vol. 18, no. 4, August 1993, pp. 395-410.
  3. For more information, talk to John Lui.

LUI9804

Title: Distributed Virtual Environment
Supervisor: Dr. John C.S. Lui
Number of students: 2
Project Type: CS or CE
Prerequisite: love system programming, motivated, willing to learn new things, independent and has intellectual curiosity.
Project description: The aim of this project is to use the ''Distributed Interactive Virtual Environment'' (DIVE) toolkit to build a simple multiple servers, multiple clients virtual environment game. (For more information about DIVE, please refer to http://www.sics.se/dce/dive/dive.html)

The goal is to allow multiple servers to dynamically manage several attach clients so that different clients can view a subset of the virtual worlds. To do this project, an understanding of the DIVE is essential. After we understand the DIVE toolkit, then we can use the partitioning algorithm proposed by my graduate students to design a simple multiple servers multiple clients games.

References:
  1. For more information, talk to John Lui, M.F. Chan, T.S. Tam and Oldfield So.