|
|
Good-bye and Thank You from eNews Daily!
 Like the banana boat heading off into the sunset, we at INFORMS must say good-bye to San Diego and our wonderful meeting attendees. We hope you enjoyed the coverage of important sessions, feature articles, session change information, and photos. We look forward to seeing you next year at the INFORMS 2010 Annual Meeting in Austin, Texas, November 7-10. For coverage of Wednesday's sessions and events, be sure to check your in-box next week for the 2009 INFORMS Annual Meeting Daily eNews wrap-up issue. Farewell!
|
|
Global Supply Chain Management Education
by Jessica Ng
 There is a great potential for academic and commercial growth in addressing supply chain innovation on a global level. As the Director of the Massachusetts Institute of Technology's Engineering System Division and the MIT Center for Transportation and Logistics (CTL), Dr. Yossi Sheffi has done just that: he helped in the development of an education structure that allows for the learning and sharing of supply chain knowledge internationally. Founded in 1973 as an interdisciplinary unit in the MIT School of Engineering, MIT's own CTL is a leader in research in transportation, logistics, and supply chain management. The program combines research, education, and outreach to the industry. In 2003, MIT created the Zaragoza Logistics Center (ZLC) in Zaragoza, Spain. ZLC is a logistic hub of high-speed railway, highway exchanges, and an airport accommodating the world's largest cargo planes. In placing the ZLC in the midst of a complex distribution center where over 250 companies operate, MIT successfully put a university in the middle of a large-scale industrial laboratory. MIT has developed several higher-level education programs in conjunction with the ZLC with the intention of demonstrating the power of a strong relationship between academia, industry, and the government. The ZLC has met great successes, being acknowledged as a Center of Excellence in Integrated Logistics and identified as a locus in Spain for scientific and technology research in supply chain management. In a similar effort in Latin America, the Center for Latin-American Logistics Innovation has met similar success, achieving national recognition as a Center of Excellence in Bogota, Columbia. The research focus of these academic institutions within the industrial setting has proven quite effective in establishing relationships and long-term contracts with commercial companies and government agencies. The combination of full-time faculty and high-level full-time professionals, as well as the involvement of other regional universities, has helped these programs achieve success furthering supply chain innovation.
|
|
Operational and Contractual Decisions in Inventory and Supply Chain Models
by Seda Tepe
Jing Zhu, the first speaker in Wednesday's session on operational and contractual decisions in inventory and supply chain models, analyzed the effects of mergers in an oligopolistic market. He presented a case where merging firms take advantage of mergers but their dependent partners do not benefit. Zhu continued his analyses with two different models. Based on his results, upstream mergers benefit both merging suppliers and the dependent retailers; however, downstream mergers benefit all the firms except for the dependent suppliers.
Jen-Yi Chen focused on a manufacturer selling goods to two newsvendors who compete in pricing and quantity recourse. Chen used the mattress industry as an example of where the retailers can sell different brands. The competition between the retailers is of the leader-follower type, and intra-trading is allowed. In this setting, the manufacturer can design the procurement contracts so that he can manage the effects of competition. This model is designed to have two stages. At the contracting stage, the manufacturer offers the contract price, and retailers sequentially decide advance purchase quantities. In the recourse stage, the manufacturer announces the spot price, the selling retailer sets the intra-trading price, and retailers simultaneously determine the trading quantity, spot purchasing quantity, and sales quantity. As a result of his analysis, Chen concluded that flexible manufacturers prefer simultaneous-move and no-trading retailers, and less flexible manufacturers who have a significantly high production cost at the second stage prefer leader-follower and intra-trading retailers.
The next speaker, Arnab Bisi stated that known demand distribution assumption is not realistic for new innovative products (i.e., fashion goods, video CDs, computers) as well as the products with volatile demand. For example, O, The Oprah Magazine is the most successful magazine in decades; after two issues, its sales averaged 1.5 million, which is more than the total of three other popular magazines. However, it is hard to predict sales before the launch of the magazine. In this study, Bisi aims to find the exact optimal policies that are tractable or computable in a censored newsvendor problem. After examining different demand distributions, Weibull distribution is shown to be the only distribution having scalable optimal policies. Bisi recommends extension of this study with different distributions to determine if exact optimal solutions exist.
Yimin Yu, the final speaker of this session, presented a study on capacity sharing and cost allocation among independent firms in the presence of congestion. Unlike the previous presentations, Yu's work is on queuing systems. Capacity sharing is the fulfillment of demand that arises from multiple sources from a single facility; airlines sharing maintenance facilities or hospitals sharing diagnostic equipment are examples of capacity sharing. Firms sharing capacity decide on capacity levels to minimize delay costs and capacity investment costs subject to service-level constraints. They consider both full information and incomplete information settings. With incomplete information, firms always have the incentive to underreport the unit delay cost. Based on the analysis presented, capacity sharing under incomplete information may be beneficial for specific parameter settings. The benefit of capacity sharing can be significant depending on cost and information sharing.
|
|
Operations Research in Healthcare Modeling
by Yichuan Ding
In Wednesday's session on modeling and analysis in healthcare systems, sponsored by the Health Applications Section and chaired by Dr. Bo Zeng of the University of South Florida, the four presenters shared the results of their recent research in healthcare applications. .
 The first speaker, Dr. Ayten Türkcan from Purdue University, recently worked in designing efficient chemotherapy operations plans for cancer treatment. Cancer is the second-leading cause of death in the world. One of the most effective therapies for cancer is chemotherapy; however, chemotherapy resources are limited and are in high demand, and the time and dose of chemotherapy has a crucial impact on its efficacy. The complexity of the operations planning of chemotherapy calls for the use of operations research methods.
Türkan and her colleagues used an integer programming method to model this problem. Their model took into account the resource limit, cyclic nature, and treatment delays of chemotherapy. The sophisticated analysis and application of optimization methods led to a near-optimal solution to the problem, which improved the efficiency of chemotherapy and reduced the treatment delays according to their results.
 The other speakers included Dr. Shengyong Wang of the University of Akron, Dr. Anahita Hassanzadeh, and Dr. Laila Cure of the University of South Florida. They introduced interesting applications of operations models in hospital inpatient bed capacity planning, dynamic ambulance relocation, and evaluation of patient safety interventions, respectively. In sum, as one of the speaker said, "Operations research has changed the delivery of healthcare in all aspects."
|
|
E-Commerce Strategy and Mechanism Design
by Seda Tepe
 Imagine you want to buy the movie Transformers and you search for it online. There are some legitimate channels like Amazon, Barnes & Noble, or Walmart (with positive cost); however, there are piracy channels like Mininova or http://www.xunlei.com (with zero cost). In this setting, it is hard to make decisions as a consumer, a legitimate channel, or a policy maker. Xianjun Geng of the University of Texas at Dallas presented a study on how to compete with piracy using a multichannel sequential search approach. In this study, legitimate and piracy channels are allowed to differ in product availability, product quality, search cost, and price distributions. On the other hand, consumers differ in their search costs; for some consumers searching is fun (no cost), but for others it is not. The proposed model aims to address several questions: What is the optimal channel choice for the consumer given a search history? How does cross-channel competition affect in-channel competition and pricing? What controls are effective in fighting piracy? A key observation of Geng's analysis is that retailer response to the piracy threat depends on in-channel competition intensity. In addition, the piracy threat does not put downward pressure on retailer pricing. This study can be extended by incorporating aspects like learning while searching, or renting versus piracy.
The second speaker of this session presented a study focusing on relational contracts and reputation mechanisms. Anjana Susarla of the University of Washington examines the form of procurement contracts with the possibility of future interactions rather than focus on prior relationships alone. Because real-life procurement contracts are not one-shot, it is important to incorporate this aspect and to measure the expectation of the future - not the shadow of the past. This study contributes the literature by incorporating the form of rent seeking and reputation within a relationship. Moreover, in the empirical analysis, Susarla provides rich measures for reputation capital. She enriches her theoretical results with examples from outsourced information technology services.
The final speaker, Bin Gu of the University of Texas at Austin, discussed online and local competition on brick-and-mortar (B&M) businesses with an application on a U.S. bank. A B&M business refers to a company that possesses a building for operations. In this study, data on individual customer depository balance from a B&M bank are used to address the impacts of competition from online and local banks. Although the local banks do not have a significant competition effect, the online banks create a threat on the B&M bank. Gu proposes strategies that can diminish the risk on the B&M bank.
|
|
Optimization/Stochastic Programming Session
by Yichuan Ding
Stochastic programming has been a focused interest in both the optimization and applied probability communities; perhaps it is because of its wide range of applications, or maybe it is because of the great challenges it presents. In Wednesday's session on robust and adaptive optimization, four speakers introduced their recent work in optimization and stochastic programming.
For example, Dr. Vineet Goyal of the Massachusetts Institute of Technology introduced his recent finding that a fixed robust solution provides a 2-approximation for a two-stage mixed-integer stochastic programming problem and an adaptive programming problem when the constrained set is of a certain type, like a hypercube. The audience found this result intriguing because the original two-stage stochastic programming problem is intractable, and Goyal's conclusion relies on fairly general assumptions.
 Dr. Zizhuo Wang of Stanford University discussed how he, along with Professors Peter Glynn and Yinyu Ye, investigated a new robust optimization model.. The classical robust model usually assumes a distribution with a fixed mean and variance. However, in the model Wang proposed, the parameters of the distribution could be more flexible compared with the classical model, but the distribution had to be consistent with the observed data, in that given this distribution, the observed data occur with a large enough probability. More interestingly, Wang claimed that this likelihood robust optimization problem can be solved in polynomial time and has interesting applications in well-known problems like the newsvendor problem and the portfolio selection problem.
The methodology research on optimization and stochastic programming has significantly improved the operation efficiency in many practical areas, and the emergence of new problems in modern management science motivates the methodology researchers to look for new meaningful models and solution techniques. The research in fundamental methodology has played an important role in improving the real-life operational efficiency, as well as research, in those more practical areas.
|
|
E-Business/Commerce I by Jessica Ng
As the amount of information on the Internet continues to grow at unprecedented levels, so does the opportunity for advancement in the field of Web analytics. The study of Web data trends can aid in better decision making and in the development of e-business/commerce strategies.
Facebook, the global leader of the social networking industry, has recently passed the 300-million-user mark. The opportunities for profit from online advertising are tremendous with such a large user base; in fact, Facebook's online advertisements are a main source of the site's revenues. The site follows a targeted-ad business model; that is, the site shows ads that cater to a user's public profile information and browsing patterns. This idea of highly targeted ads and consumers acting as marketers has proven quite effective, as seen in a small-scale case study done by Arash Farsi of Carnegie Mellon University. However, the idea of the targeted-ad business model engulfs a number of user privacy issues. Interestingly, Farsi found that the majority of users within his study were willing to sacrifice their basic privacy to use a free Facebook with targeted, personalized ads, whereas less than 10% would prefer to use a fee-based Facebook site with no advertisements at all. Based on this strategy, Facebook has placed more value on the effectiveness of its campaigns rather than the potentially negative social ramifications.
In his Ph.D. thesis, Dennis Ahrholdt of the University of Hamburg addressed the huge potentials in the online retailing, or "e-tailing," of physical goods that have gone unused. As an e-tailer of his own store, Ahrholdt recognized the use of the Internet as a search channel rather than a transacting channel. To back up his point, he mentioned the low conversion rates of those that browse small e-tailer sites to those that become an actual purchaser at those sites. Ahrholdt hypothesized that one major component within the traditional e-tailer business model is missing: trust. Ahrholdt introduces the need for a "trust model" to act as a driving force behind the intention to transact. Specifically, he notes that the use of "IT signals" provides the needed trust incentive to improve the sales performance of e-tailers. These IT signals can be simple and can come in the form of a user-friendly layout, inclusion of a seal of approval, multimedia illustrations, and risk-free payment methods. Within the scope of his studies, he found that e-tailers using these IT signals have seen higher purchasing rates. The e-tailer business model can grow tremendously by recognizing the need for trust reinforcement.
|
|
Thank You to our eNews Writers
Yichuan Ding
 Yichuan Ding is currently in his third year of doctoral study in the Department of Management Science and Engineering at Stanford University. His advisor is Dr. Stefanos Zenios, a professor at the Stanford Graduate School of Business. Yichaun's research interest lies broadly in the area of operations research, including optimization, game theory, and queue theory, with a focus on healthcare applications. His work has been published in journals such as Mathematics of Operations Research and Optimization and Engineering, and has been or will be published in conference proceedings such as the ACM-SIAM Symposium on Discrete Algorithms (SODA) and the First International Workshop on Mobile Entity Localization and Tracking in GPS-less Environments (MELT 2008).
Recently, Yichuan and his advisor developed an incentive-based multiqueue model, which will hopefully improve organ allocation efficiency. Yichaun has also worked in robust and stochastic optimization, along with Stanford faculty members Dr. Amin Saberi and Dr. Yinyu Ye. Prior to coming to the United States, he obtained an M.Math degree at University of Waterloo, Canada, where he was supervised by Dr. Henry Wolkowicz, a professor in the Department of Combinatorics and Optimization. Yichuan, a citizen of China, holds a B.S. degree in honored mixed class at Zhejiang University.
Yichuan has been actively involved in extracurricular and social activities. As the Vice President of Chinese Student and Scholars Association, he organized many social activities, such as mixers, dances, and art performances, for the sake of improving the quality of life for Stanford students. Last but not least, Yichuan has been a comedic actor, a singer, a dancer, and a chorus conductor in his past performance experience.
Seda Tepe
 Seda Tepe is a recent graduate of the Master of Science program in Industrial and Systems Engineering at the University of Minnesota (UMN). During her studies at UMN, she worked as a research assistant and teaching assistant. She conducted research in inventory management and stochastic modeling. She also worked as a teaching assistant for the graduate-level courses Project Management and Engineering Optimization.
Prior to joining UMN, Seda obtained an M.Sc. degree in industrial engineering from Koç University, Turkey, where she was a teaching assistant for three courses: Probability, Decision Analysis, and Operations Management. For her master's thesis, she studied the impacts of advance demand information in multiclass production-inventory systems.
Seda is interested in research and applications of supply chain management, revenue management, mathematical programming, and decision analysis.
|
|
Please Take Our eNews Survey
We hope you enjoyed the 2009 Annual Meeting eNews Daily. The four issues delivered to your email in-boxes on Sunday, Monday, Tuesday, and Wednesday mornings were designed to bring you a recap of the previous day's activities and preview of the coming day's activities. We hope you enjoyed the news accounts of key sessions, features articles, session change information, upcoming events, and photos. In an effort to better serve future meeting attendee needs, please tell us what you liked and did not like about our eNews Daily. Go here and fill out our survey. We thank you for your opinion!
|
|
|
|
|
|
