21 August 2017 in Management Science
Currently available medication for treating many chronic diseases is often effective only for a subgroup of patients, and biomarkers accurately assessing whether an individual belongs to this subgroup typically do not exist. In such settings, physicians learn about the effectiveness of a drug primarily through experimentation—i.e., by initiating treatment and monitoring the patient’s response. Precise guidelines for discontinuing treatment are often lacking or left entirely to the physician’s discretion. We introduce a framework for developing adaptive, personalized treatments for such chronic diseases. Our model is based on a continuous-time, multi-armed bandit setting where drug effectiveness is assessed by aggregating information from several channels: by continuously monitoring the state of the patient, but also by (not) observing the occurrence of particular infrequent health events, such as relapses or disease flare-ups. Recognizing that the timing and severity of such events provide critical information for treatment decisions is a key point of departure in our framework compared with typical (bandit) models used in healthcare. We show that the model can be analyzed in closed form for several settings of interest, resulting in optimal policies that are intuitive and may have practical appeal. We illustrate the effectiveness of the methodology by developing a set of efficient treatment policies for multiple sclerosis, which we then use to benchmark several existing treatment guidelines.The online appendix is available at https://doi.org/10.1287/mnsc.2017.2793 .This paper was accepted by Noah Gans, stochastic models and simulation.
1 October 2018 in Management Science
Firms often finance their inventory through debt and subsequently sell it to generate profits and service the debt. Pricing of products is consequently driven by inventory and debt servicing considerations. We show that limited liability under debt induces sellers to charge higher prices and to discount products at a slower pace. We find that these distortions result in revenue losses that compound over time, leading to some form of performance spiral down. We quantify the extent to which these inefficiencies can be mitigated by practical debt contract terms that emerge as natural remedies from our analysis, and find debt amortization or financial covenants to be the most effective, followed by debt relief and early repayment options. The online appendix is available at https://doi.org/10.1287/mnsc.2017.2862 . This paper was accepted by Serguei Netessine, operations management.
1 March 2020 in Manufacturing & Service Operations Management
Problem definition: Loyalty programs (LPs) introduce a new currency—the points—through which customers transact with firms. Such points represent a promise for future service, and their monetary value thus counts as a liability on the issuing firms’ balance sheets. Consequently, adjusting the value of points has a first-order effect on profitability and performance and emerges as a core operating decision. We study the problem of optimally setting the points’ value in view of their associated liabilities. Academic/practical relevance: Firms across numerous industries increasingly utilize LPs. The sheer magnitude of LPs coupled with recent changes in accounting rules have turned the associated liabilities into significant balance-sheet items, amounting to billions of dollars. Managers (from chief financial officers to chief marketing officers) struggle with the problem of adjusting the points’ value in view of these liabilities. Academic work is primarily aimed at understanding LPs as marketing tools, without studying the liability angle. Methodology: We develop a multiperiod model and use dynamic programming techniques and comparative statics analysis. Results: We show that the optimal policies depend on a new financial metric, given by the sum of the firm’s realized cash flows and outstanding deferred revenue, which we refer to as the profit potential. The total value of loyalty points is set to hit a particular target, which increases with the profit potential. We find that loyalty programs can act as buffers against uncertainty, with the value of points increasing (decreasing) under strong (weak) operating performance and increasing with uncertainty. Managerial implications: Setting the point values and adjusting operating decisions in view of LP liabilities should be done by tracking the firm’s profit potential. Loyalty programs can act as hedging tools against uncertainty in future operating performance, which provides a new rationale for their existence, even in the absence of competition.
1 July 2021 in Management Science
We consider a system with an evolving state that can be stopped at any time by a decision maker (DM), yielding a state-dependent reward. The DM does not observe the state except for a limited number of monitoring times, which he must choose, in conjunction with a suitable stopping policy, to maximize his reward. Dealing with these types of stopping problems, which arise in a variety of applications from healthcare to finance, often requires excessive amounts of data for calibration purposes and prohibitive computational resources. To overcome these challenges, we propose a robust optimization approach, whereby adaptive uncertainty sets capture the information acquired through monitoring. We consider two versions of the problem—static and dynamic—depending on how the monitoring times are chosen. We show that, under certain conditions, the same worst-case reward is achievable under either static or dynamic monitoring. This allows recovering the optimal dynamic monitoring policy by resolving static versions of the problem. We discuss cases when the static problem becomes tractable and highlight conditions when monitoring at equidistant times is optimal. Lastly, we showcase our framework in the context of a healthcare problem (monitoring heart-transplant patients for cardiac allograft vasculopathy), where we design optimal monitoring policies that substantially improve over the status quo recommendations. This paper was accepted by Chung Piaw Teo, optimization.
1 June 2021 in Management Science
Many operational settings share the following three features: (i) a centralized planning system allocates tasks to workers or service providers, (ii) the providers generate value by completing the tasks, and (iii) the completion of tasks influences the providers’ welfare. In such cases, the planning system’s allocations often entail trade-offs between the service providers’ welfare and the total value that is generated (or that accrues to the system itself), and concern arises that allocations that are good under one metric may perform poorly under the other. We propose a broad framework for quantifying the magnitude of value losses when allocations are restricted to satisfy certain desirable guarantees to the service providers. We consider a general class of guarantees that includes many considerations of practical interest arising (e.g., in the design of sustainable two-sided markets) in workforce welfare and compensation, or in sourcing and payments in supply chains, among other application domains. We derive tight bounds on the relative value loss and show that this loss is limited for any restriction included in our general class. Our analysis shows that when many providers are present, the largest losses are driven by fairness considerations, whereas when few providers are present, they are driven by the heterogeneity in the providers’ effectiveness to generate value; when providers are perfectly homogenous, the losses never exceed 50%. We study additional loss drivers and find that less variability in the value of jobs and a more balanced supply-demand ratio may lead to larger losses. Lastly, we demonstrate numerically using both real-world and synthetic data that the loss can be small in several cases of practical interest. This paper was accepted by Chung Piaw Teo, optimization.
1 January 2015 in SSRN Electronic Journal
This paper is the first to study the dynamic management of loyalty programs adopting a holistic view of their interaction with the operations, marketing, and fi