Events & Seminars

Newcomers are important for the survival of online communities, but their contributions often receive negative reactions and comments from established members. Online communities realize that such negativity can take a toll on newcomers and harm the creation of user­-generated content. We study a novel intervention aimed at reducing hostility toward newcomers: a “newcomer nudge” that informs community members when they are interacting with a newcomer’s post and asks them to be more lenient toward its creator. Taking advantage of granular data from a large deal­-sharing community and a natural experiment, we use a difference-­in­-differences approach and find robust evidence that the newcomer nudge induced members to write 46% more responses per day with 10% fewer negative words during the first two days after a deal was published. Our results show that the nudge­-induced change in behavior toward newcomers increased newcomer retention. However, we also observe that before the nudge, newcomers’ second posts received more net votes (upvotes minus downvotes) than their first posts. After the nudge, newcomers’ subsequent posts were less popular than their first posts, which indicates that the nudge interrupted newcomers’ learning curve by suppressing helpful feedback.

In recent years, we have seen a surge of interest in supply chain visibility. Under this paradigm, decision- makers are able to trace the real-time data (e.g., stock level, resource allocation flow) along the entire supply chain so that they can identify the decision-making bottlenecks and take actions more efficiently. Motivated by the Gaze Heuristic, we propose a target-based online planning framework to deal with real-time resource allocation problems in both stationary and nonstationary environments. Leveraging on the Blackwell's Approachability Theorem and Online Convex Optimization tools, we characterize the near-optimal performance guarantee of our online solution in comparison with the offline optimal solution, and explore the properties of different allocation policies.

We use synthetic and real data from various industries, from supply chain planning in manufacturing, to resource deployment in ride-sharing markets, to examine the impact and value of these real-time solutions in practice: (1) we present a new insight into the impact of supply chain visibility on the capacity configuration in the capacity pooling system. Our results show that the pooling system does not need to hold any safety stock to deliver the required demand fulfillment service if real-time allocation with full visibility is utilized, when the number of customers is sufficiently large in the system; (2) we study a real-time ride-matching problem in the ride-sourcing context, with multi-objectives (e.g., service quality, revenue) to be considered. We develop a new technique that can be used to choose the weight adaptively over time, based on real-time tracking of the gaps in attained performance and a set of performance targets. Our results show that the real-time matching policy could potentially contribute to the long-term sustainability and reputation of the ride-sourcing platform by dispatching more orders to drivers with higher service quality, without sacrificing the short-term platform revenue.

Video games represent the largest and fastest-growing segment of the entertainment industry, which involves 3 billion gamers and garners $180 billion annually. Despite its popularity in practice, it has received limited attention from the operations community. Managing product monetization and engagement presents unique challenges due to the characteristics of gaming platforms, where players and the gaming platform have repeated (and endogenously controlled) interactions. In this talk, we describe a body of work that provides the first analytical results for this emerging market. In the first part, we discuss a prevailing selling mechanism in online gaming known as a loot box. A loot box can be viewed as a random bundle of virtual items, whose contents are not revealed until after purchase. We consider how to optimally price and design loot boxes from the perspective of a revenue-maximizing video game company, and provide insights on customer surplus and protection under such selling strategies. In the second part, we consider how to manage player engagement in a game where players are repeatedly matched to compete against one another. Players have different skill levels which affect the outcomes of matches, and the win-loss record influences their willingness to remain engaged. Leveraging optimization and real data, we provide insights on how engagement may increase with optimal matching policies, adding AI bots, and providing a pay-to-win feature.

Operations management research could provide great insights into healthcare operations. In this seminar, I will describe two of my most recent projects related to healthcare operations. In the first paper (with Rath and Coleman), we study collaborative care model to treat patients who suffer from Diabetes and Depression. In particular, we create a model that facilitates decision support for such collaborative care. In the second paper (with Bhatia), using inpatient discharge data from hospitals in California between 2008-2016, we create a metric for standardization of healthcare services delivered to patients. Leveraging our standardization metric, we examine the impact of a hospital’s healthcare service standardization on its cost, quality, and variation in both cost and quality of service.

Many Operations Management problems involve two-stage decision-making and hence are computationally difficult to be solved in general. In this work, we solve a class of two-stage distributionally robust optimization problems which have the property of supermodularity. We exploit the explicit worst-case expectation of supermodular functions and derive the worst-case distribution for the robust counterpart. This enables us to develop an efficient method to obtain an exact optimal solution of these two-stage problems. We also show that the optimal scenario-wise segregated affine decision rule returns the same optimal value in our setting. Further, we provide a necessary and sufficient condition for checking whether any given two-stage optimization problem has the supermodularity property. We apply this framework to several classic problems, including the multi-item newsvendor problem, the facility location design problem, the lot-sizing problem on a network, the appointment scheduling problem and the assemble-to-order problem. While these problems are typically computationally challenging, they can be solved efficiently using our approach.

Service level requirement is an important measure of service quality in the real-life business. A big challenge for the companies is to appropriately harness the resources to meet their target service levels for customers. Companies gain competitive advantages by optimizing (1) the capacity level of pooled resources in anticipation of random demand of multiple customers and (2) the capacity allocation to fulfill customer demands after demand realization.

We present a general framework to study this two-stage resource allocation problem when customers require individual and possibly different service levels. Our modeling framework generalizes and unifies many existing models in the literature. We propose a simple randomized rationing policy for any fixed feasible capacity level. Our main result is the optimality of this Max-Weighted-Service policy for very general service-level constraints, including Type-I and Type-II constraints and beyond. The result follows from a semi- infinite linear programming formulation of the problem and its dual. We also prove the optimality of priority policies for a large class of problems when the set of feasible fulfilled demands is a polymatroid. Moreover, with a slight change in one step of the Max-Weighted-Service policy, it is also optimal when there is differentiated allocation cost from resources to demands. This is based on joint work with Jiashuo Jiang and Jiawei Zhang from NYU stern.

Financial asset price movement impacts product demand, and thus influences the pricing and production decisions of a firm. We develop and solve a general model that integrates pricing, production, and financial risk hedging decisions for firms of newsvendor type. We find that in general, the presence of hedging reduces the optimal price; it also reduces the optimal service level when the asset price positively impacts the product demand (“asset price benefits demand”), while it may increase the optimal service level by a small margin when the impact is negative (“asset price hurts demand”). We construct the mean-variance efficient frontier that characterizes the risk-return trade-off and quantify the risk reduction achieved by the hedging decision. Our numerical case study using real data of Ford Motor Company shows that the markdowns in pricing and service levels are small under our model, and the hedging decision can substantially reduce risk without materially decreasing operational profit.

On-demand service platforms have its gig workers to use self-set nonbinding performance goals to regulate their effort and overcome potential self-control problems. To examine the effect of such self-goal setting mechanisms, we build a behavioral model, derive theoretic results and testable hypotheses, and conduct a field experiment on a large gig platform of food deliveries. The model incorporates the reference-dependent utility theory of goal setting into the two-self framework of self-control. Our model analysis finds that individual workers' optimal self-set goal may exhibit a spectrum of difficulty level, ranging from trivially to impossibly achievable, depending on their reference-dependent utility coefficients and their self-control cost; and that their effort is always higher with a properly set goal than the no-goal benchmark, although the difference is significant only when both the reference-dependent utility coefficients and the self-control cost are sufficiently large. Our experiment data confirms heterogeneous treatment effects: While the average treatment effect is insignificant, a causal tree algorithm identifies a sub-group of population whose effort significantly increases under the goal- setting treatments. Our study compares the two common types of performance metrics for goal setting, the number of completed orders versus the total revenue. Both our model and experiment data suggest that the two types of goals lead to equal effort improvement but different attainment probabilities. In particular, the goal attainment rate is lower in the revenue-goal treatment than the quantity-goal treatment because workers tent to set excessively high revenue goal. Our study demonstrates the efficacy and the limitations of self-goal setting mechanisms, and yields two important managerial implications. First, there exists a reasonably sized population for target marketing of the self-goal setting mechanisms; second, platforms would better encourage the use of order-quantity goals instead of revenue-goals for higher attainment rates.

We consider the problem of optimal fleet sizing in a vehicle sharing system. Vehicles are available for short-term rental and are accessible from multiple locations. The size of the fleet must account not only for the nominal load and for the randomness in demand and rental duration but also for the randomness in the number of vehicles that are available at each location due to vehicle roaming (vehicles not returning to the same location from which they were picked up). We model the system as a closed queueing network and obtain a closed form approximation of the optimal fleet size (the minimum number of vehicles needed to meet a target service level). The approximation is remarkably accurate and highly interpretable with buffer capacity expressed in terms of three explicit terms that can be interpreted as follows: (1) standard buffer capacity that is protection against randomness in demand and rental times; (2) buffer capacity that is protection against vehicle roaming; and (3) a correction term. Our analysis reveals important differences between the optimal sizing of standard queueing systems and that of systems where servers roam.

Agriculture is changing in many ways. This talk gives an overview of these changes, with a particular focus on agricultural supply chains. Like many supply chains around the world, agricultural supply chains are subject to digital disruption in a variety of interesting ways. I will outline what these disruptions mean for agriculture today and make some projections for the future. New Zealand case studies will be presented. Further, I will also discuss new agricultural technologies and precision agriculture and what they mean for research in this important area. Ideas for future research will be discussed throughout the talk. At the end I will give some broader perspectives on research publishing.