Service Engineering and Management

sem-title.jpg

Major pillars of the Hong Kong economy are related to services such as finance, professional services, medicine, education and logistics. Those service systems are complex systems in which specific arrangements of people and technologies take actions that provide value for others. Systems are designed and built to provide and sustain services, yet because of their complexity and size, operations do not always go smoothly, and all interactions and results cannot be anticipated. As a result, systems engineers are trained to develop quantitative decision-making tools and methodologies for smooth, agile and resilient operations in data-intensive service systems such as finance, healthcare, and logistics.

Effective Control Strategies against Healthcare Associated Infections via Optimization Approaches 
C.H. Cheng
Healthcare associated infections (HAIs) are the disease that people acquire inside a healthcare setting such as hospital. This type of infections poses a serious threat to public health. For instance, there were about 722,000 cases of HAIs reported by U.S. hospitals in 2011. About 10% of HAI cases resulted in death (Magill et al. [1]). Another example was the 2003 SARS epidemic causing hundreds of deaths around the world. It was later traced back to an outbreak in a hospital in Hong Kong (Lee et al. [2] and Hung [3]).
 
Modern healthcare systems are trying to reduce the risk of HAIs. Among many control strategies against infectious diseases, targeted immunization is a common and effective  approach for preventing infections among people. It requires identifying potential “super spreaders” and preventing them from being the sources that spread the diseases over human contact networks, with the goal of reducing the impact of epidemic outbreaks as far as possible. Except for direct protection for targeted individuals, targeted immunization also provides indirect protection for people who are in the same contact networks as the targeted individuals. These people  may be infants and pregnant women who are not suitable for getting vaccinated themselves. Moreover, when vaccines are scarce, possibly due to limited budgets or shortages of  supply, allocating resource optimally is essentially important for eveloping efficient intervention strategies. 
 
In this work, we plan to develop effective and efficient control strategies of target immunization against HAIs. This work leverages on the PI’s previous project on developing a hospital risk management system with indoor application of the RFID technology. This technology tracks the location and movement of human and medical equipment and monitors real-time ward events. We have conducted a four-month pilot study of  our system in a public hospital in Hong Kong. Based on the  data collected, we plan to conduct further investigation for developing network-based control strategies for minimizing HAI outbreaks. Specifically, several activities are involved to achieve this goal:
  • dynamic human contact network construction based on human tracking data in a hospital for characterizing personto-person interaction in a healthcare setting, 
  • concise formulation of targeted immunization as an outbreak minimization problem under network diffusion models, 
  • investigation on the equivalence between the outbreak minimization problem and the influence maximization problem which is extensively studied in the field of social influence analysis, 
  • solution to the outbreak minimization problem using optimization approaches, which is applicable to real-world scenarios, 
  • experimental evaluations on real data and comparison to the existing methods to verify the performance and scalability.  
Our work is unique for four reasons. First, we focus on the indoor application of targeted immunization in a relatively closed community against HAIs, while most existing work generally  deals with outdoor cases. Second, human interaction pattern in a healthcare setting differs from that in a large population due to the highly hierarchical and modular structure of a hospital and distinct roles of patients and caregivers. Third, we solve the outbreak minimization problem with optimization approaches instead of greedy methods used in most existing work. Finally, our work makes contribution to the study on infectious disease control as well as the study on influence maximization by improving the solution efficiency and effectiveness. 
 

Financial Digital Library
J. Yu, C.C. Yang and W. Lam
The Financial Digital Library being developed contains annual reports, financial news articles, and government documents that allows users from different places to access and search for the information they need based on concept space. We have a collection of annual reports from 249 Hong Kong public firms, real-time stock quotes, and a set of agents to support technical and fundamental analysis. We have also conducted a series of studies on how an electronic filing system can improve transparency of financial information transmission in Hong Kong.

Integration of OLAP and Multidimensional Inter Transaction Mining
J. Yu
Today’s markets are much more competitive and dynamic than ever before. Business enterprises prosper or fail according to the sophistication and speed of their information systems, and their ability to analyse and synthesize information using those systems. Integration of On-Line Analytical Processing (OLAP) and data mining is a promising direction since it facilitates interactive exploratory data analysis. The objective of this project aims at integrating OLAP and multidimensional inter-transaction data mining for large financial multidimensional databases.

Integration of Renewable Energy Resources with the Sustainable Water System in Hong Kong: Models and Algorithms
C.H. Cheng & Neng Fan
Energy is required to extract, collect, pump and deliver water for human consumption and industrial use, and also for wastewater treatment before its reuse or its return to the environment. For example, the US alone uses at least 521 million MWh a year for water-related operations, and this is equivalent to 13% of the country’s total electricity consumption. According to US Environmental Protection Agency, the massive energy consumption in water/wastewater treatment and delivery add over 45 million tons of greenhouse gases annually.
To reduce further damage to the environment and at the same time to meet water demands sustainably, one of the most effective approaches relies on usage of renewable energy resources. However, these resources, such as solar, wind, etc., are highly unpredictable and intermittent due to weather conditions. Meanwhile, water demands are also highly uncertain depending on the future population growth and climate changes.
A large-scale integration of renewable energy into water systems faces the uncertainties in these two infrastructures. To overcome these challenges, we propose a research to develop optimization-based models and algorithms for policy makers to integrate renewable energy as a major energy source, and to optimally operate sustainable water systems. To achieve these goals, we plan to conduct the following research activities:
(1) Quantifying the uncertainties of water demands and renewable energy resources through data mining and data analytics.
(2) Designing multistage stochastic programming models and algorithms for renewable integration to water systems, and also for management and operations of the water/wastewater treatment facilities.
(3) Evaluating the proposed techniques for practical test in a real water system.
Different from previous studies, this research is unique. First, instead of only considering the energy efficiency in a water system, we concentrate on large renewable integration for both planning, and management and operations of the water system. Second, the proposed multistage stochastic programming model will have the capability to incorporate the uncertainties because of renewable integration. Finally, the success of this project will be a practical and methodological evidence for the concept of onsite energy generation to promote the adoption of renewable energy in other industries.

Knowledge Discovery
W. Lam, H. Meng and J. Yu
This project focuses on automated or semi-automated learning from data and texts, and the transformation of learned theories into some knowledge representation formalisms. We expect to develop the theory and techniques for partial or full automation of the time-consuming process of expert knowledge elicitation through automatic knowledge discovery or learning from data. We aim not only at the accuracy and effectiveness of the learned information, but also at improving the level and depth of knowledge discovered.

Network Epidemiology Modeling of Dynamic Human Behaviors for Controlling Hospital Acquired Diseases
C.H. Cheng & Dorbin Tobun Ng
The SARS epidemic in 2003 was traced back to an outbreak in a hospital in Hong Kong. In 2013, a Frenchman died of a SARS-like virus. He got infected while he was sharing a common hospital room with an infective. To ensure public health, we need a further understanding of human interaction dynamics in a hospital environment and a deeper investigation of nosocomial infections for effective and efficient control strategies.
Time and space play an important role in shaping human behaviors, inferring human interactions, and influencing epidemic spread of infectious diseases in a healthcare institute. Our research aims at enhancing the capability of tracking the movement of people and tracing the infection of a hospital-acquired disease for each individual in order to monitor and control the epidemic spread of nosocomial infections. To achieve this goal, we plan to conduct the following research activities:
(1) network models of dynamic human behaviors for an enhanced spatial-temporal analysis to determine human interaction patterns and contacts in a relatively close community,
(2) epidemic models to track and trace nosocomial infections based on the time-varying contact networks to increase our understanding of transmission mechanisms in hospital,
(3) traceability analysis of tracking the mobility and infection for each individual to identify risk behaviors by linking both the network and epidemic models,
(4) effective and efficient control strategies against epidemic spread under different scenarios to take advantage of the network structure and disease dynamics.
This research is unique. First, since our focus is on the movement of people, our methods are very different from those used to track products in a supply chain. Further, we plan to conduct the traceability analysis of the dynamic interactions and infections of people at individual level for network epidemiology study in hospital. This differs from either the popular modeling of a static network or the traditional epidemiology study for a large region. Finally, many spatial-temporal analyses and epidemiology models are developed for outdoor applications while this work deals with indoor activities.