Technical Sessions – Monday July 11

Pap smears administered to women with proper frequency can reduce incidents
of cervical cancer by 60% by signaling a need for further tests. However,
the Pap smear is subject to sensitivity and specificity errors. To account for
this uncertainty, we propose a partially observable Markov decision process
(POMDP) in order to formulate an optimal policy for the screening and treatment
of cervical cancer. This policy considers the risk factors of the patient
as well as test results and observations on disease progression. The model is
applied to patient data from a healthcare provider in Colombia.
2 - An Inferential Problem of Recurrent Breast Cancer in
Survival Analysis
Chichang Chang, Dept of Applied Information Sciences,
Chungshan Medical University, 110, Sec. 1, Chien-Kuo N. Rd.,
40201, Taichung, Taiwan, threec@csmu.edu.tw
All cancers are classified according to the stage so that therapies may be tailored
to the particular situation. Detection of asymptomatic recurrences is associated
with prolonged survival and from the initial detection. This paper applied
Bayesian reference analysis to produce objective, model-based, posterior
distributions, to an inferential problem of recurrent breast cancer in survival
analysis. The sampling distribution is modeled through a proportional intensity
homogeneous Poisson process. And, attempts to improve surveillance and
assessment of recurrent status could improve outcome.
3 - An Empirical Study of the Electronic Medical Record
Project in Taiwan
Yunglin Cheng, Applied Information Sciences, Chungshan
Medical University, Taiwan, hdst520@hotmail.com, Chichang
Chang
This research is going to offer cautions of Electronic Medical Record security
and measure risks successfully by building a frame of information security
management in hospital. This research developes an alerted model by two-step
preference decision of parameter-free method which can reflect risky attitude
and decision behavior of the participants completely.The result shows that our
research meet with prior literature which evidences that preferences between
risky prospects are not linear in probabilities, and the methods of measuring
risk was thought to be more reliable then Riskit model.
4 - OR in healthcare: how to win friends and influence people
Don Campbell, General Medicine, SouthernHealth, Monash
Medical Centre, 246 Clayton Road, 3168, Clayton, Victoria,
Australia, donald.campbell@monash.edu, Keith Stockman
Whilst it has been clear for years that OR has great potential in health it’s influence
has been limited. Asking "What are the problems?" and "How can OR
help?" yields little. Expanded close collaboration between OR and Health is
needed. OR must more effectively embed itself in the system and learn how
to influence decisions. It must better understand the problems that health faces
and the realities of making change. Education of healthcare workers in OR is
needed along with the injecton of break through ideas and thinking. How this
might be achieved will be discussed.
� TB-17
Tuesday, 11:00-12:30
Meeting Room 214
Network Design and Routing
Stream: Network Optimisation and Telecommunications
Invited session
Chair: Bernard Fortz, Département d’Informatique, Université Libre
de Bruxelles, CP 210/01, Bld du Triomphe, 1050, Bruxelles,
Belgium, bfortz@euro-online.org
1 - Models for Optimal Survivable Routing with a Minimum
Number of Hops: Comparing Disaggregated with Aggregated
Models
Luis Gouveia, DEIO, University of Lisbon, Campo Grande,
Bloco C6, 1749-016, Lisbon, Portugal, legouveia@fc.ul.pt,
Pedro Patrício, Amaro de Sousa
IFORS 2011 - Melbourne TB-18
Given an undirected network with link capacities and a set of commodities
with known demands, this paper addresses the problem of determining D hopconstrained
node disjoint paths for each commodity while minimizing the average
or the maximum number of hops. These paths are defined according to
two survivability mechanisms: Path Diversity and Path Protection. We present
two classes of ILP models, disaggregated and aggregated, for both problems,
study the relationship between their linear programming relaxations and compare
their effectiveness through a set of computational experiments.
2 - K-edge Survivability in Ring Networks
Young-Soo Myung, Dept. of Business Administration, Dankook
University, 330-714, Cheonan, Korea, Republic Of,
myung@dankook.ac.kr
Given an undirected network with a set of source-sink pairs, we are assumed
to get a benefit if a pair of source and sink nodes are connected. The k-edge
survivability of a network is defined as the total benefit secured after arbitrarily
selected k edges are destroyed. The problem of computing k-edge survivability
is known to be NP-hard and has applications of evaluating the survivability
or vulnerability of a network. In this paper, we consider the k-edge survivability
problem restricted to an undirected ring network and develop an exact
algorithm to solve it in polynomial time.
3 - Solving Two Facility Network Design Problem - A Polyhedral
Approach
Faiz Hamid, Information Technology & Systems, Indian Institute
of Management, Lucknow, India, FPM Office, IIM Lucknow,
Prabandh Nagar, Off-Sitapur Road, 226013, Lucknow, Uttar
Pradesh, India, faiz@iiml.ac.in, Yogesh Agarwal
The paper studies the problem of designing telecommunication networks using
transmission facilities of two different capacities. A new family of facets is
identified and also a new approach for computing the facets is introduced based
on polarity theory. The approach has been tested on several randomly generated
networks. Computational results show that 3-partition facets reduce the integrality
gap, compared to that provided by 2-partition facets, by approximately
30-60%. Also there is a substantial reduction in the size of branch-and-bound
tree.
4 - FIPP p-Cycles Based Resilient Network Design for
MPLS
Richard Harris, School of Engineering and Adv Tech, Massey
University, Private Bag 11 222, Manawatu Campus, 4474,
Palmerston North, Manawatu, New Zealand,
r.harris@massey.ac.nz, Jing Zhang
An extension of p-cycles, Failure-Independent Path-Protecting (FIPP) p-cycles
are an efficient protection architecture for networks. They use properties of pcycles:
mesh-like efficiency, ring-like speed, and concurrently provide shared,
failure independent end-to-end protection for working paths. We study FIPP
p-cycle based resilient networking by two models for FIPP p-cycles: FIPP-
SCP and FIPP-DRS. We consider joint capacity allocation design using the
FIPP-SCP model for use with Multi Protocol Label Switching (MPLS) in Next
Generation Networks. We discuss several comparative case studies.
� TB-18
Tuesday, 11:00-12:30
Meeting Room 215
DEA- Keynote & Tutorial
Stream: Data Envelopment Analysis
Invited session
Chair: Ali Emrouznejad, Aston Business School, Aston University,
B4 7ET, Birmingham, United Kingdom, a.emrouznejad@aston.ac.uk
1 - A Systematic Process for Measuring Efficiency of Decision
Making Unites
Ali Emrouznejad, Aston Business School, Aston University, B4
7ET, Birmingham, United Kingdom,
a.emrouznejad@aston.ac.uk
47