1. Your full name,
address and e-mail address:
Wuyi Yue
Prof., Department of Information
Science and Systems Engineering
Faculty of Science and
Engineering
Director, Institute of Intelligent
Information and Communications Technology (IICT)
Konan University
8-9-1 Okamoto, Higashinada-ku
Kobe 658-8501, JAPAN
e-mail: yue@konan-u.ac.jp
2. Your highest
degree, awarding institution and year
3. How many
research papers have you published (including papers accepted for publication)?
How many of them in the field of optimization?
4. Your
research interests:
My research interests include
queueing theory and its application to system modeling, performance analysis,
performance evaluation and optimal
resource allocation for mobile wireless communication
networks, multimedia communication networks, traffic systems, stochastic systems
and information systems
to maximize the network utilization and minimize the call blocking probability
and transmission delay; optimization
problems in the Markov decision processes; systems engineering and operations
research.
5. Some of your
most representative papers or books:
Books:
[1] Wireless Computer
Communication, McGraw-Hill Press, Editors: N. Seshagiri and Aram Akopov,
pp.87-98, 1994.
[2] Wuyi Yue and yutaka
Matsumoto, Performance Analysis of Multichannel and Multi-Traffic on Wireless
Communication Networks, Kluwer Academic Publishers, p324, 2002.
Journal Papers:
[1] Wuyi Yue, “Analytical
Methods to Calculate the Performance of a Cellular Mobile Radio Communication
System with Hybrid Channel Assignment,” IEEE Transactions on Vehicular
Technology, Vol. 40, No. 2, pp. 453-460, 1991.
[2] Wuyi Yue, “The Effect of
Capture on Performance of Multichannel Slotted ALOHA Systems,” IEEE Transactions
on Communications, Vol. 39, No. 6, pp. 818-822, 1991.
[3] Wuyi Yue and Yutaka
Matsumoto, “Output and Delay Processes in a Slotted ALOHA Multichannel Packet
Radio Network with Capture,” Journal of Probability in the Engineering and
Informational Sciences, Vol. 6, No. 4, pp. 471-493, 1992.
[4] Wuyi Yue, Chifa Ku, Jiayu
Nie and Pingxian Hu, “System modeling and optimization of a Port System with
Different Ship's Kinds,” Journal of
Systems Science and Mathematical Science, Vol. 6, No. 4, pp. 296-311, 1993.
[5] Wuyi Yue and Yutaka
Matsumoto, “Performance Analysis of CSMA/CD with Slotted Multiple Channel on
Radio LANs,” Journal of Computer Communications, Vol. 16, No. 10, pp. 637-644,
1993.
[6] Wuyi Yue and Yutaka
Matsumoto, “Probability Distribution of Delay in Cellular Mobile Networks with
Hand-off.” IEICE Transactions on Communications, Vol. E79-A, No. 7, pp.
1011-1020, 1996.
[7] Ting-Jie Lu, Wuyi Yue and
Toshiharu Hasegawa, “A Mutual Overflow System with Simultaneous Occupation of
Resources,” Journal of the Operations Research Society of Japan, Vol. 41, No. 1,
pp. 81-90, 1998.
[8] Wuyi Yue and Yutaka
Matsumoto, “Output and Delay of Multi-channel Slotted ALOHA Systems for
Integrated Voice and Data Transmission,” Journal of Telecommunication Systems,
Vol. 13, No. 2-4, pp. 147-165, 2000.
[9] Wuyi Yue and Yutaka
Matsumoto, “Output and Delay Process Analysis for Slotted CDMA Wireless
Communication Networks with Integrated Voice/Data transmission,” IEEE Journal on
Selected Areas in Communications, Vol. 18, No. 7, pp. 1245-1253, 2000.
[10] Wuyi Yue and Yutaka
Matsumoto, “Exact Analysis of Multi-traffic Wireless Communication Networks with
Reserved and Nonreserved Multi-channel,” IEICE Transactions on Communications,
Vol. E-84-B, No. 4, pp.786-794, 2001.
[11] Wuyi Yue and Yutaka
Matsumoto, “A New Effective Analysis for Wireless CSMA/CA LANs Supporting
Real-Time Voice and Data Services,” IEICE Transactions on Fundamentals of
Electronics, Communications and Computer Sciences, Vol. E-84-A, No. 7, pp.
1660-1669, 2001.
[12] Wuyi Yue and Yutaka
Matsumoto, “Performance Analysis of CSMA/CA Protocol for High-Speed Wireless
LANs,” Journal of International Transactions in Operational Research, Vol. 9,
No. 1, pp. 85-96, 2002.
[13] Qiying Hu and Wuyi Yue,
“Optimal Replacement of a System According to a Semi-Markov Decision Process in
a Semi-Markov Environment,” Journal of Optimization Methods and Software, Vol.
18, No. 2, pp. 181-196, 2003.
[14] Qiying Hu, Jianyong Liu
and Wuyi Yue, “Necessary Conditions for Continuous Time Markov Decision
Processes with Expected Discounted Total Rewards,” International Journal of Pure
and Applied Mathematics, Vol. 7, No. 2, pp. 147-176, 2003.
[15] Wuyi Yue and Koichi
Hatogai, “Performance Evaluation of Wireless LAN on Multimedia Information
Communication Environments,” Journal of Dynamics of Continuous, Discrete and
Impulsive Systems, Special Issue: Dynamic Systems in Communications Networks,
Vol. 10, No. 4, pp. 561-580, 2003.
[16] Qiying Hu and Wuyi Yue,
“Analysis for Some Properties of Discrete Time Markov Decision Processes,”
Journal of Optimization, Vol. 52, No. 4-5, pp. 495-505, 2003.
[17] Wuyi Yue, Jifa Gu and
Xijin Tang, “A Performance Evaluation Index System for Multimedia Communication
Networks and Forecasting for Web-Based Network Traffic,” Journal of Systems
Science and Systems Engineering, Vol. 13, No. 1, pp. 78-97, 2004.
[18] Wuyi Yue and Koji
Miyazaki, “Optimal Channel Assignment in Wireless Communication Networks with
Distance and Frequency Interferences,” Journal of Computer Communications, Vol.
27, pp. 1661-1669, 2004.
[19] Qiying Hu and Wuyi Yue,
“Two New Optimal Models for Controlling Discrete Event Systems,” Journal of
Industrial Management and Optimization, Vol. 1, No. 1, pp. 65-80, 2005.
6. Please
describe your major contributions in optimization
(1)
Optimization
Problems in Network Designs
To
satisfy the huge service demand and improve the system performance for network
optimization, efficient traffic access protocols
and analytical methods must be provided. Especially,
modern
communication
networks to support several different traffic types and
several
different
hopes to retransmit have
become so complex that intuition alone is not sufficient to predict their
performance and optimization.
Mathematical
modeling and performance optimization have come to play an important role in the
workings of communication networks.
Queueing networks and Markov
chains are commonly used for the performance and reliability evaluation of
communication system, information systems and stochastic systems for optimal
designs of the systems. On the other hand, the optimal design and development of
communication system, information systems and stochastic systems require not
only average performance measures as throughput, transmission delay and call
blocking probability but also higher moments of traffic departures and traffic
delay. One purpose of our researches, therefore, is to offer detailed exact and
approximate analytical solution methods and techniques using queueing theory to
model and analyze the complex communication systems, information systems and
stochastic systems with procedures of multiple random access schemes and
reliably evaluate the performance of the systems to help the optimal network
designs like maximize
the network utilization or minimize the call blocking probability and data
transmission delay.
(2)
Optimal Resource Allocation
Since
radio bandwidth is still a limited resource in the mobile wireless communication
industry, providing optimal utilization of the networks has become a pressing
dilemma for the mobile communication research community. Resource allocation
such as channel assignment in communication networks is very important. Our
researches include to present efficient algorithms and stochastic models for
optimizing resource allocation in communication networks. Fixed channel
assignment in communication networks is a significant combinatorial optimization
problem that must be solved. The search space for channel assignment grows
exponentially with the network size. Finding an optimal solution is a difficult
problem. However, a good solution of channel assignments can reduce interference
to communication and reduce the response to users requests. Therefore, it is
worthwhile to make an effort to find good approximate solution to the optimal
solution. The goal of our research is to achieve optimal channel assignment
taking a centralized view in the whole network. We suggest approaches combining
several important heuristics. Our experimental results show that our algorithm
improves over known approaches.
We
also present stochastic models and analyses for optimizing resource allocation
by considering the
complicated elements in
communication networks. Also, cost analysis of such communication networks helps
in the designing of efficient routing protocols and call admission control
schemes. We present methods to get the optimal resource allocation to minimize
the total cost in the networks system.
(3)
Capacity Optimization in Communication Networks
Traffic engineering in communication networks is a process of controlling
traffic demand in a network so as to optimize resource utilization and network
performance. There are two forms of traffic engineering: online planning and
offline planning. Offline traffic engineering simultaneously examines each
channel's resource constraints and the requirements of each Local Service
Provider (LSP) to provide global calculations and solutions for the
communication networks in a centralized view. We develop stochastic traffic
engineering optimization models focusing on both the random traffic demand and
the network performance optimization, such as loss rate and network risk
management. By using mean-variance approach and conditional Value-at-Risk, we
study the impact of risk averseness on the network profit function and introduce
a penalty cost in the optimization model for network bandwidth allocation. We
give numerical results to show the impacts of relationship between the revenue
and the cost, loss rate constraint and risk averseness on the network
performance.
(4)
Optimal Problems
in Markov Decision Processes
In our researches, we present
new methodologies for discounted Markov decision processes (MDPs). This is to
reduce the scale of MDPs models based on action reduction and state
decomposition. The idea of the action reduction is that an action can be
eliminated if any policy use it would not be optimal. The main ideas about the
state decomposition are that we decompose the state space of the model into
several sub-spaces. In each subspace, we know an optimal policy or we can easily
solve the MDPs model. Thus, the original MDPs model is decomposed into several
smaller MDPs models. The purpose to reduce the scale of MDPs model is mainly to
separate the case with finite optimal value from the cases with positive or
negative infinite optimal value, and then we can just study the case with finite
optimal value.
By using these methodologies
above, we study discrete time MDPs models and continuous time MDPs models with
the discounted criterion under the necessary condition. We also present new
optimal control problems for discrete event systems and study them by using
these methodologies and study optimal replacement problems in stochastic
environments.
7. What are the
most interesting unsolved problems in the optimization branch you are working
on:
The
three entities that are the main important points to traffic analysis and design
are external traffic load, engineered resources, and observed performance. Given
the amount of resources and the traffic load, we should evaluate the system
performance such as channel utilization, throughput, packet delay, call
blocking, and high moments about traffic characters. The random service demands
and multimedia service demands complicate mobile
wireless
communication systems to process traffic loads.
In
mobile
wireless communication networks, several multimedia sources are statistically
multiplexed over the network links with user mobility and user's random
multi-traffic demand. Efficient channel access protocols must be employed to
utilize the limited spectrum among all the mobiles efficiently. Wireless
communication system designers need methods for the quantification of system
design factors such as performance and reliability. We usually want to know
which network protocol gives the best delay throughput characteristic under
specified conditions, how coefficients of variation of the packet delay and the
packet interdeparture time change with the offered traffic load, what size
buffers must be employed by a network to keep the probability of buffer overflow
below a particular value, and what is the maximum number of voice calls that can
be accepted by a network in order to keep the voice packet transfer delay within
reasonable bounded, what is the optimal network
performance and
so on.
We
can answer these and other related questions for the wireless communication
networks by developing queueing models and Markovian models, and then analyze
these models to obtain such performance measures and offer numerical results for
network
performance optimization.
Probabilistic and statistical methods are commonly employed for the purpose of
performance and reliability evaluation.
The
most direct method for performance evaluation is based on actual measurement of
the system under study. However during the design phase the system is not
available for such experiments and yet performance of a given design needs to be
predicted to verify that it meets design requirements and to carry out necessary
trade-offs. Hence abstract models are necessary for performance prediction of
designs.
On the other hand, mobile
wireless communications with the demand for voice, Internet, and multimedia
services for fixed and mobile user have a bright future. We would like to
suggest several important future research topics for wireless communication
networks. We offer some discussions as to the kinds of wireless communication
systems that will emerge over the next few year and our considerations on the
model and methodology for these wireless communication systems.
Performance
optimization problems arising in the study of telecommunication networks
basically consist of two components: traffic modeling and queuing analysis.
Traffic is the driving force behind all telecommunications activity, and its
models are of crucial importance for evaluating network performance. Probability
models for traffic streams are needed to the extent that they allow prediction
of system performance measures to a reasonable degree of accuracy. However,
traffic modeling and queuing analysis based on traditional Poisson arrival and
exponential call holding time assumptions, which have performed successfully in
the design and analysis of circuit-switched telephone networks, are no longer
valid for high-speed packet-switched networks carrying LAN-to-LAN data,
compressed video, and so on with considering the use mobility and multi-hop.
To
evaluate, improve and optimize such system’s performance, we attempt to provide
efficient performance analysis methods, efficient queuing analysis techniques
and tools to design, engineer, and dimension networks. In most cases where very
complex queuing models are encountered, we believe that our performance analysis
methods will be successfully employed and sometimes preferred to analysis due to
the intractability of such models and will be useful for performance
optimization of next generation
wireless
communication network systems.
8. What
kinds of topics excite your research interests?
Due to its widespread
applications, network optimization is an important subfield within the broad
field of optimization. Our
researches are
still related to communication network optimization problems: maximum network
utilization, shortest-path problem, minimum-cost single-commodity flow
problem and minimum-cost multicommodity flow problem. These applications are
intended to illustrate a range of problem contexts and to be suggestive of how
network optimization problems arise in practice.
9. How did you
develop these interests? What would you say is one of the most interesting
topics you have studied?
Still another
problem is the routing of traffic on the network. In a resource-sharing
communication environment, transmission route choice models or transmission
assignment models aim to describe traffic resources on communication networks,
which operate at known switches. The route choice models or traffic assignment
models are also resource control problems with a waiting phenomenon: traffic
experiences a waiting time for the switch of the line on which traffic is
chosen.
The modeling of
such communication networks is taken to be the minimization of expected waiting
and transmission time, or the expected total generalized
cost if waiting
times and transmission times may have different weights.
Given potential sites for network node location, traffic data, and
available link types and their cost, the lowest-cost local access network
configuration must be obtained considering optimum variables. That being a
network with an optimum number of network nodes and their locations, optimum set
of links interconnecting the network nodes, their capacities, routing paths,
subject to constraints on delay, throughput, reliability and link capacity.
10. To round off
the interview, what are some highlights of your career?
In our
future researches, we
will give three research
results related to performance optimization methods for the resource control
problems in wireless
multi-traffic and multi-hop communication
networks. The first research introduces a novel concept of backlog balancing and
demonstrates its application to network resource control
and congestion control by presenting a rate-based resource control
algorithm for next
generation
wireless
communication
networks. The aim of such resource control
models is to maximize the network utilization for achieving high throughput with
tolerable delay for each access point.
The second
research results provide a performance optimization analysis of virtual route
networks using queuing theory for which a pacing window resource control
mechanism is employed with an input queue included. Messages prevented from
entering the network are stored in the input queue in a first-come first-serve
manner.
Both cases of
finite and infinite capacity of buffer are considered. The results show that
although the average number of messages in the network is higher, when the input
queue delay is taken into consideration. The overall
performance of the system is better than that of the other systems.
The third
research proposes a resource control
scheme for data packet traffic to avoid network congestion and to obtain the
maximum throughput. The resource control
algorithm of transmission control protocol determines the packet transmission
rate using the congestion window size, which is adjusted to obtain a large
throughput without network congestion and buffer over
resource of the
receiver. The resource control
algorithm decreases the window size when network congestion is detected.
Fairness, stability, and optimality of the proposed method are discussed with
respect
to the performance of the system.