[{"id":19660,"title":"Analytics and Operations Young Scholar Workshop 2024","permalink":"https:\/\/bschool.nus.edu.sg\/biz-events\/event\/analytics-and-operations-young-scholar-workshop-2024\/","category":"Seminars and talks","event_dept":{"value":"analytics-operations","label":"Analytics & Operations"},"event_sec_dept":false,"event_details":{"event_start_date":"1 March 2024","event_end_date":"1 March 2024","event_start_time":"10:00 am","event_end_time":"6:05 pm","event_dress_code":"NA"},"event_loc":{"eve_address_selection":"5","eve_location_1":{"eve_org":"NUS Business School","eve_build":"Mochtar Riady Building","eve_room":"3-2","eve_add":"15 Kent Ridge Drive","eve_count":"Singapore","eve_copos":119245,"eve_map_url":"https:\/\/goo.gl\/maps\/Q1kyjwxHNE22"},"eve_location_2":{"eve_org":"Shaw Foundation Alumni House","eve_build":"","eve_room":"Clove and Lemongrass Room Level 2","eve_add":"11 Kent Ridge Drive","eve_count":"Singapore","eve_copos":119244,"eve_map_url":"https:\/\/goo.gl\/maps\/docgThkDWFxKdb9c7"},"eve_location_3":{"eve_org":"Hon Sui Sen Memorial Library Auditorium","eve_build":"","eve_room":"","eve_add":"1 Hon Sui Sen Drive","eve_count":"Singapore","eve_copos":117588,"eve_map_url":"https:\/\/goo.gl\/maps\/NJjWK4RMpC92"},"eve_location_4":{"eve_org":"NUSS Kent Ridge Guild House","eve_build":"","eve_room":"Dalvey Room","eve_add":"9 Kent Ridge Drive","eve_count":"Singapore","eve_copos":119241,"eve_map_url":"https:\/\/goo.gl\/maps\/nXn2Luh96pH2"},"eve_location_5":{"eve_org":"Institute of Data Science","eve_build":"Innovation 4.0","eve_room":"I4-01-03 (Level 1 Seminar Room)","eve_add":"3 Research Link","eve_count":"Singapore","eve_copos":"117602","eve_map_url":"https:\/\/goo.gl\/maps\/i1xocvvDh27QUXem7"},"eve_location_6":{"eve_org":"","eve_build":"","eve_room":"","eve_add":"","eve_count":"","eve_copos":"","eve_map_url":""},"eve_location_7":""},"event_introduction":"","event_short_intro":"","event_topic":null,"event_banner":false,"event_external_url":"","event_registration_details":{"event_registration_form":false,"event_registration_message":"","event_registration_deadline":null,"eve_registration_url":"","event_form":"","event_registration_ack":""},"event_speaker":[{"event_speaker_name":"Yifan Feng","event_speaker_designation":"Assistant Professor, NUS Bussiness School","event_speaker_affiliation":"","event_speaker_picture":false,"event_speaker_url":"https:\/\/bizfaculty.nus.edu.sg\/faculty-details\/?profId=632","event_speaker_introduction":""},{"event_speaker_name":"Kimon Drakopoulos","event_speaker_designation":"Associate Professor of Data Sciences and Operations, USC Marshall School of Business","event_speaker_affiliation":"","event_speaker_picture":false,"event_speaker_url":"https:\/\/www.kimondrakopoulos.com\/","event_speaker_introduction":""},{"event_speaker_name":"Yvonne Huijun Zhu","event_speaker_designation":"PhD Student, NUS IORA","event_speaker_affiliation":"","event_speaker_picture":false,"event_speaker_url":"https:\/\/iora.nus.edu.sg\/people-p\/zhu-huijun\/","event_speaker_introduction":""},{"event_speaker_name":"Zhi Chen","event_speaker_designation":"Assistant Professor, NUS Business School","event_speaker_affiliation":"","event_speaker_picture":false,"event_speaker_url":"https:\/\/bizfaculty.nus.edu.sg\/faculty-details\/?profId=633","event_speaker_introduction":""},{"event_speaker_name":"Pascale Crama","event_speaker_designation":"Professor of Operations Management, SMU Lee Kong Chian School of Business","event_speaker_affiliation":"","event_speaker_picture":false,"event_speaker_url":"https:\/\/business.smu.edu.sg\/faculty\/profile\/6666\/crama-pascale","event_speaker_introduction":""},{"event_speaker_name":"Zhaoxuan Wei","event_speaker_designation":"PhD Student, NUS IORA","event_speaker_affiliation":"","event_speaker_picture":false,"event_speaker_url":"https:\/\/iora.nus.edu.sg\/people-p\/wei-zhaoxuan\/","event_speaker_introduction":""},{"event_speaker_name":"Long Zhao","event_speaker_designation":"Assistant Professor, NUS Business School","event_speaker_affiliation":"","event_speaker_picture":false,"event_speaker_url":"https:\/\/bizfaculty.nus.edu.sg\/faculty-details\/?profId=586","event_speaker_introduction":""},{"event_speaker_name":"Ilia Tsetlin","event_speaker_designation":"Professor of Decision Sciences, INSEAD","event_speaker_affiliation":"","event_speaker_picture":false,"event_speaker_url":"https:\/\/www.insead.edu\/faculty\/ilia-tsetlin","event_speaker_introduction":""},{"event_speaker_name":"Lan Wu","event_speaker_designation":"PhD Student, NUS Business School","event_speaker_affiliation":"","event_speaker_picture":false,"event_speaker_url":"https:\/\/bschool.nus.edu.sg\/analytics-operations\/wp-content\/uploads\/sites\/73\/2022\/08\/CV_WuLan.pdf","event_speaker_introduction":""}],"event_agenda":false,"event_photo_gallery":false,"event_presentations":false,"event_custom_heading":[{"event_custom_title":"Paper 1","event_custom_details":"

Presenter<\/strong>: Yifan Feng, Assistant Professor, NUS Bussiness School<\/p>\n

Discussant<\/strong>: Kimon Drakopoulos, Associate Professor of Data
\nSciences and Operations, USC Marshall School of Business<\/p>\n

Learning to Rank under Strategic “Brush Wars”<\/strong><\/p>\n

Abstract:
\nWe consider a dynamic learning and ranking problem of a digital
\nplatform. Uninformed of the products’ intrinsic qualities, the platform
\nstrives to design a sequential ranking policy that learns from historical
\ntraffic data while accounting for potential manipulation by sellers to
\ninflate their performances, which we refer to as “brushing.” Are there
\neffective yet simple ranking algorithms to combat manipulative ranking
\nbrushing?
\nWe provide a positive answer by proposing a simple ranking algorithm,
\ntermed Experiment-Then-Commit (ETC). We study the sellers’
\nstrategic responses to the ranking algorithm by formulating an
\nN-player-T-period “brush war” game. We first show that the dynamic
\ngame admits a static reduction through a dominant-strategy argument.
\nThen for every fixed N, we characterize the game’s asymptotic
\nbehavior when T is large. We show the nonexistence of pure strategy
\nequilibria, shedding light on the possibility of efficiency loss. However,
\nfor a large market where N is infinite, a different pattern emerges. We
\nformulate a novel non-atomic game with a continuum of sellers and
\ncharacterize a “self-reinforcing” market equilibrium. Under this
\nequilibrium, the seller’s brushing amount increases in the product’s
\nquality. In other words, the sellers’ strategic responses “reinforce”
\ncomplete learning of the platform. As a result, ETC can be highly
\neffective even under sellers’ manipulative ranking brushing. We also
\ndiscuss the managerial implications.<\/p>\n"},{"event_custom_title":"Paper 2","event_custom_details":"

Presenter<\/strong>: Yvonne Huijun Zhu, PhD Student, NUS IORA<\/p>\n

On the Value of Flexibility in Adaptive Experiments<\/strong><\/p>\n

Abstract:
\nAdaptive experimentation refers to the practice of changing (improving)
\nthe experiment configurations on the fly based on observed data.
\nWhile it can potentially increase sample efficiency, it potentially needs
\nstrong flexibility as an experimenter and thus may be difficult to
\nimplement in practice.
\nTo explore the interplay between sample efficiency and flexibility in
\nadaptive experiments, we formulate a Bayesian sequential hypothesis
\ntesting problem. The goal is to minimize the sum of sampling and
\npenalty costs under different levels of flexibility. Here, flexibility is
\ndefined as the number of times one can change the experiments. We
\nshow that there is a notable gap between “No Flexibility” (where
\nexperiments cannot be changed) and “Full Flexibility” (where
\nexperiments can be freely changed adaptively). Nevertheless, the gap
\ncan be (asymptotically) closed by just allowing to change the
\nexperiment once. That is achieved by an experimentation rule we
\ndevelop, termed Nested SPRT, which is both simple to derive in closed
\nform and easy to implement.<\/p>\n"},{"event_custom_title":"Paper 3","event_custom_details":"

Presenter<\/strong>: Zhi Chen, Assistant Professor, NUS Business School<\/p>\n

Discussant<\/strong>: Pascale Crama, Professor of Operations Management,
\nSMU Lee Kong Chian School of Business<\/p>\n

The Value of Private Feedback in Trial-and-error Innovation<\/strong>
\nContests<\/strong><\/p>\n

Abstract:
\nFirms have increasingly turned to innovation contests as a means of
\nprocuring complex industrial innovations from their supplier base. At
\nthe start of the contest, the buyer announces the innovation challenge
\nand the award in the form of a valuable supply contract. To address the
\nchallenge, it is common that suppliers adopt the trial-and-error
\napproach to develop solutions, and then submit them to the buyer for
\nevaluation as the contest progresses. In some contests, the buyer
\nprivately reveals the interim performance of developed solutions to the
\nsuppliers during the trial-and-error process (private feedback), but in
\nothers, the buyer withholds such information (no feedback). Motivated
\nby such divergent practices, we seek to understand whether providing
\nprivate feedback (or not) results in higher profits for the buyer. We find
\nthat when there is no urgent need for faster time-to-market, providing
\nprivate feedback is more profitable for the buyer when either there are
\nmany evaluation rounds or when the cost of a trial is high. We further
\nuncover two sources of the value of private feedback: it incentivizes
\nsuppliers to conduct more trials (\u201cquantity effect”) in a more efficient
\nway (\u201cquality effect”). In contrast, no feedback results in higher profits
\nfor the buyer when there are few evaluation rounds and when the cost
\nof a trial is low. On the contrary, when the faster time-to-market is a
\nmajor concern for the buyer, we show that private feedback becomes
\nless attractive, and the buyer prefers suppliers to conduct parallel trials
\n(no feedback) particularly when the cost of a trial is very low or high, or
\nwhen the competition is intense with a large number of suppliers. Our
\nresults have direct managerial implications and help shed light on
\nvarious feedback policies used in practice.<\/p>\n"},{"event_custom_title":"Paper 4","event_custom_details":"

Presenter<\/strong>: Zhaoxuan Wei, PhD Student, NUS IORA<\/p>\n

Partial Backorder Inventory System: Asymptotic Optimality and<\/strong>
\nDemand Learning<\/strong><\/p>\n

Abstract:
\nWe develop a unified stochastic inventory model that not only captures
\nthe impatience feature for unmet demands but also integrates the
\nclassic backlogging and lost-sales inventory models. For such model,
\nwhen both demand and patience distributions are known, we establish
\nthe uniform (asymptotic) optimality of the base-stock policy. While the
\nbacklogged demands become unobservable (demand is partially
\nobservable and patience is unobservable), we introduce a novel policy
\nfamily that operates without backlogged demands information, and
\nprove this proposed policy can approach the cost efficiency of the
\noptimal policy in the system when the demand and patience
\ndistributions are known. Further we extend our analysis to an online
\ninventory control problem in which precise data on demand and
\npatience distributions are unobservable and only sales are observable
\nby developing a UCB-type algorithm that yields a near-optimal policy.
\nThe regret bounds given by the algorithm are provably tight within the
\nplanning horizon, and are comparable to the state-of-the-art results in
\nthe literature, even in the face of partial and biased observations and
\nweaker system ergodicity.<\/p>\n"},{"event_custom_title":"Paper 5","event_custom_details":"

Presenter<\/strong>: Long Zhao, Assistant Professor, NUS Business School<\/p>\n

Discussant<\/strong>: Ilia Tsetlin, Professor of Decision Sciences, INSEAD<\/p>\n

Predicting Tail Quantiles Through Aggregation of Medians: Model<\/strong>
\nand Analysis<\/strong><\/p>\n

Abstract:
\nQuantile forecasts are essential inputs for decision making under
\nuncertainty. The most useful quantiles are from the tails since tails
\nprovide rich information about the underlying uncertainty. However,
\ndirectly predicting tail quantiles is challenging for human experts and
\neven quantitative models. On the other hand, non-tail quantiles such
\nas medians are easier to predict directly but are less useful in
\nuncertainty assessment. Motivated by this mismatch, we study how a
\ndecision maker can predict tail quantiles through aggregation of
\nnon-tail quantile forecasts (such as medians). Intuitively, we want to
\nshift the median forecasts by the true difference between median and
\ntail quantiles. However, this task is challenging because the underlying
\nrandomness is unknown, and the median forecasts are subject to
\nbiases and noises. Our method involves aggregation of median
\nforecasts to achieve a substantial reduction of noise and leveraging
\npast data to correct the biases and estimate the shift. We evaluate our
\nproposed method using the M5 uncertainty competition submissions
\nand find that our method outperforms established benchmarks in the
\nliterature. We also offer a theoretical understanding of why our method
\nperforms well empirically.<\/p>\n"},{"event_custom_title":"Paper 6","event_custom_details":"

Presenter<\/strong>: Lan Wu, PhD Student, NUS Business School<\/p>\n

Mitigating the Spiral Down Effect: Online Learning under<\/strong>
\nMixed-fare Structure<\/strong><\/p>\n

Abstract:
\nIn the dynamic landscape of airline revenue management, accurately
\npredicting customer behavior and adjusting ticket-selling strategies in
\nthe face of fluctuating demand is critical. Traditional models frequently
\noverlook intricate customer buy-down patterns, resulting in a
\ndetrimental downward spiral of revenue decline. This study combines
\nthe adjusted fare concept with the Bayes Selector algorithm to create
\nan innovative approach for optimizing airline ticket-selling strategies
\nunder mixed fare structure. By dynamically adjusting fare structures
\nbased on purchase probabilities and expected profits, and utilizing
\nprobabilistic estimates and learning capabilities of the Bayes Selector
\nalgorithm, this integrated methodology enables airlines to adaptively
\nrefine their seat allocation strategy for maximum profitability.
\nOur findings show that a constant regret is attainable for the problem
\nunder the (independent) differentiated product demand setting. More
\ninterestingly, we establish a logarithmic regret bound in the case of
\nmixed fare structure. To validate these results, we provide a simple
\nanalysis of how estimation errors in the probabilities of the types (of
\ncustomers) is managed in the online seat allocation algorithm.
\nFurthermore, we illustrate the practical application of our approach
\nthrough a real-world airline scenario. This study concludes that by
\nlearning and adapting to the nuances of customer behavior, airlines
\ncan significantly enhance their revenue management capabilities,
\nleading to more accurate demand estimation and seat allocation
\nstrategy in alignment with the existing environment.<\/p>\n"}],"event_enquiry_details":{"event_enq_full_name":"","event_enq_department":"","event_enq_email":"","event_enq_telephone":"","event_enq_website":""}}]