Operations Research
December 1, 2006
Melbourne, Australia
|
|
Recent Advances in Operations Research December 1, 2006 Melbourne, Australia |
|
| Welcome | News | Important dates | Program | Registration | Hotels | Venue | Contact | Sponsors | FAQs |
We have a very tight schedule, so please ... help us implement it as planned!
Hopefully this is a very good approximation of the final program, but ... we may have to modify it slightly. Check this page next week for a possible update. We shall list the changes to the current program (4:45PM, Thursday, November 23, 2006).
Schedule:
Abstracts:
Gad Abraham (UoM): Modelling the Flow of Elective-Surgery Patients.
The goal of hospital elective waiting-list managers is to maximise elective occupancy while minimising cancellations due to urgent admissions. Patient length of stay is random, making it difficult to manually assess the impact of decisions. We present preliminary work on a tool for day-to-day elective-planning decision support. It will enable the user to make short-term predictions of patient discharges and occupancy levels. The user can explore different admission scenarios, and predict their impact on overall elective occupancy. This is useful both for maintaining a predetermined level of elective patient occupancy, and for monitoring the short term behaviour of elective admissions.
Leon Au (UoM): A different perspective towards modelling Leave without Being Seen (LWBS) patient.
Patients who arrive to the Emergency Department (ED) without being seen by a physician are known as Leave Without Being Seen (LWBS) patients. Over the years, the number of LWBS patients has increased and so patients are being turned away from receiving proper assessment and treatment. As a result, the hospital could potentially be exposed to liability and bad publicity by the media.
In this presentation, an overview of the current literature will be discussed, and recommendation for a proposal towards a different perspective in modelling LWBS patient through the hospital data warehouse. Finally, I will conclude with a discussion about system thinking.
Tristan Barnett and Alan Brown, Swinburne (UoM): Generalizing the 'Kelly' betting strategy to allow for multiple payouts.
The 'Kelly' strategy for calculating the optimal bet size for a single payout is well established. This paper provides a generalization of the 'Kelly' strategy for when multiple payouts exist. Video Poker is an example of a game where this new formulation can be applied to assist in determining an optimal bet size. A Jacks or Better machine is given as a working example. From an Operations Research perspective, the 'Kelly' strategy has also been applied to securities markets and information theory.
Leonid Churilov, Zoe Zhuang, and Ken Sikaris (Monash): Intelligent decision support for pathology ordering by general practitioners in Australia.
Pathology ordering by General Practitioners (GPs) is a significant contributor to rising health care costs both in Australia and worldwide. A thorough understanding of the nature and patterns of pathology utilization is an essential requirement for addressing the issues of appropriate pathology orderings and possible over-utilization. This presentation describes how Kohonen's Self-Organizing Maps are used to discover the most typical patterns in pathology orderings for different patient groups. Test ordering data from a pathology company in Australia is analyzed; homogenous clusters of patients with similar ordering patterns are discovered and investigated, and possible management implications are discussed. We then demonstrate how combined use of data mining and case based reasoning can facilitate intelligent decision support for pathology ordering that is both patient-oriented and deeply rooted in practical peer-group evidence. In particular, we argue that combining case-based reasoning principles that are inherently close to GPs' daily practice, and data-driven knowledge discovery mechanisms that can be applied to massive amounts of the pathology requests data routinely available at professional pathology companies, can lead to more informed evidential decision making by doctors.
Bruce Craven (UoM): Feedback model for inventories in a network
Classical EOQ models minimize the cost of inventory, while meeting a fluctuating demand, but may disregard the role of inventory in reducing the (often costly) fluctuations in production rate. An alternative model considers production and inventory together, with a feedback term to stabilize the system. This is applied to inventories in a network, which may be optimized as a whole, instead of independently optimizing each profit centre.
Cindy Hauser (UoM): Counting kangaroos and other wildlife: a balance between accuracy and cost
Monitoring is a vital component in managing an ecological system. Feedback on the current state of the system can inform the next management action, and also provide new understanding of how the system works. Given that monitoring incurs a cost and can potentially be conducted with differing levels of accuracy, ecologists have recently begun to ask: how much monitoring effort is required to achieve our goals? I will discuss the development of a state-dependent monitoring plan for a harvested red kangaroo population, and outline a new project on monitoring invasive pests.
John Hearne (RMIT): Black Rhino Translocation Policy Revisited
Between 1970 and 1990 12 African countries lost their entire populations of Black Rhino. A decade ago the population had dropped to 2400 from about one million in 1900. With high Black Rhino population densities in the Umfolozi-Hluhluwe Game Reserve it was decided to translocate some animals to new reserves on an annual basis. It was felt that the growth rates of both the source and translocated populations would increase as a result of the reduced population densities. This problem was modelled in the early 90šs and conclusions drawn. In this presentation we revisit the problem and discuss some of the management tensions that have arisen as a result of conflicting objectives. These objectives relate to the need to maximise the growth of the global population while meeting local game-viewing tourism objectives. Results to assist managers make a sensible trade-off will be presented. Some conclusions regarding the present and previous approaches will be drawn.
Kim Levy (UoM): Ready, Get Set, Switch!
We introduce the problem of real-time tracking of a moving target, such as a signal, through noisy measurements. We propose to use stochastic approximation to successively obtain better estimates of the target. The objective is to remain within acceptable bounds around the target. Our previous research has led us to develop an adaptive algorithm that maximizes the reaction time when a change of regime occurs. We will show how change detection can significantly improve the performance of the algorithm during the transitions between regimes. We present results for a new regression based hypothesis test and for the well-known Page-Hinkley test.
Dina Neiger, Leonid Churilov, Kristian Rotaru (Monash): Integrating risks in business process models with value focused process engineering.
As enterprise systems develop, the integration of various business management dimensions becomes increasingly important. However, historical disciplinary boundaries between information systems and management sciences can obstruct this path to integration. For example, risk management is generally considered as a business process within process engineering, while in the context of management sciences risk is treated as a threat to business objectives that needs to be minimized. Both views are essential for a complete approach to identifying, understanding and managing risks in order to optimally meet business requirements. In this talk, we address the need for a holistic business view of risk management in the enterprise systems space by drawing on the strengths of the respective disciplines and identifying links between their complementary views of risk, which enables us to integrate these apparently diverging views. Through the application of value-focused process engineering principles to risk management models, we develop a framework that extends the capabilities of existing enterprise systems and enables risk-oriented process management which incorporates a multi-disciplinary view of risk. The proposed framework is illustrated in the context of a critical organizational business process to demonstrate the practical application of and the resulting benefits from the use of this framework.
Moshe Sniedovich (UoM): What exactly is Info-Gap? An OR perspective.
Info-Gap is a new theory for decision making under severe uncertainty. Its claim to fame is that it is non-probabilistic in nature and thus offers an alternative to all current theories for decision making under uncertainty. The Operational Research perspective on this theory reveals that there are serious issues regarding its role and place in decision theory. I also show that the Info-Gap's uncertainty model is fundamentally flawed and consequently the solutions it generates are unlikely to be robust. The lesson to Operations Research is that it cannot take for granted that its established methods and techniques are readily available to analysts from other disciplines.
Marcus Volz (UoM): Using Networks to Optimise Costs in Underground Mine Design.
In this talk we present ways in which networks can be used to model and optimise the costs associated with underground mines. We describe two techniques for modelling an underground mine: the Steiner tree problem, and the Fermat-Weber problem. In both cases links in the network are gradient-constrained, meaning they have absolute gradient less than some upper bound. We discuss the importance of weighted edges in the networks. We conclude with an industry case study.
Rene Weiskircher (CSIRO): Simulating Trading Agents in a Network.
We present a system that simulates producer, consumer and merchant agents who trade over a network. Negotiations between the agents determine the product flows in the network. The emphasis of our research is on understanding the effects of different agent negotiation strategies on the resulting flows and the profits of the agents. Our simulation system enables us to evaluate the influences of negotiation strategies, network topology and internal agent constraints on the resulting product flows.
Jiapu Zhang (CSIRO): Energy Minimization of van der Waals (VDW) interactions.
When optimization is applied to clusters, crystals, and biomolecules, the energy minimization of VDW contacts is always considered. It is also called global optimization of Lennard-Jones (LJ) potential. This talk reports some important results that I got in Ballarat, on this problem. Its extension in structural biology and structure-based drug design is also very important; and a very important result that I found - Asn178 plays a key role for causing CJD (Creutzfeld-Jakob disease), vCJD (variant CJD), fCJD (familial/inherited CJD), FFI (fatal familial insomnia), GSS (Gerstmann-Straussler-Scheinker disease) - won an Award of CSIRO. These results are dedicated to the memory of Professor Alexander Rubinov.