An official website of the United States government
We describe a fluid model with time-varying input that approximates a multiclass many-server queue with general reneging distribution and multiple customer classes (specifically, the multiclass G/GI/N+GI queue). The system dynamics depend on the policy, which is a rule for determining when to serve a given customer class. The class of admissible control policies are those that are head-of-the-line (HL) and nonanticipating. For a sequence of many-server queues operating under admissible HL control policies and satisfying some mild asymptotic conditions, we establish a tightness result for the sequence of fluid scaled queue state descriptors and associated processes and show that limit points of such sequences are fluid model solutions almost surely. The tightness result together with the characterization of distributional limit points as fluid model solutions almost surely provides a foundation for the analysis of particular HL control policies of interest. We leverage these results to analyze a set of admissible HL control policies that we introduce, called weighted random buffer selection (WRBS), and an associated WRBS fluid model that allows multiple classes to be partially served in the fluid limit (which is in contrast to previously analyzed static priority policies).
more »
« less
A fluid approximation for a matching model with general reneging distributions
Motivated by a service platform, we study a two-sided network where heterogeneous demand (customers) and heterogeneous supply (workers) arrive randomly over time to get matched. Customers and workers arrive with a randomly sampled patience time (also known as reneging time in the literature) and are lost if forced to wait longer than that time to be matched. The system dynamics depend on the matching policy, which determines when to match a particular customer class with a particular worker class. Matches between classes use the head-of-line customer and worker from each class. Since customer and worker arrival processes can be very general counting processes, and the reneging times can be sampled from any finite mean distribution that is absolutely continuous, the state descriptor must track the age-in-system for every customer and worker waiting in order to be Markovian, as well as the time elapsed since the last arrival for every class. We develop a measure-valued fluid model that approximates the evolution of the discrete-event stochastic matching model and prove its solution is unique under a fixed matching policy. For a sequence of matching models, we establish a tightness result for the associated sequence of fluid-scaled state descriptors and show that any distributional limit point is a fluid model solution almost surely. When arrival rates are constant, we characterize the invariant states of the fluid model solution and show convergence to these invariant states as time becomes large. Finally, again when arrival rates are constant, we establish another tightness result for the sequence of fluid-scaled state descriptors distributed according to a stationary distribution and show that any subsequence converges to an invariant state. As a consequence, the fluid and time limits can be interchanged, which justifies regarding invariant states as first order approximations to stationary distributions.
more »
« less
- Award ID(s):
- 2054505
- PAR ID:
- 10515883
- Publisher / Repository:
- Springer
- Date Published:
- Journal Name:
- Queueing Systems
- Volume:
- 106
- Issue:
- 3-4
- ISSN:
- 0257-0130
- Page Range / eLocation ID:
- 199 to 238
- Subject(s) / Keyword(s):
- service platforms two-sided platform reneging fluid approximation functional limit theorems measure-valued process
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
In e-commerce, customers have an unknown patience in terms of how far down the page they are willing to scroll. In light of this, how should products be ranked? The e-commerce retailer’s problem is further complicated by the fact that the supply of each product may be limited, and that multiple customers who are interested in these products will arrive over time. In “Online Matching Frameworks Under Stochastic Rewards, Product Ranking, and Unknown Patience,” Brubach, Grammel, Ma, and Srinivasan provide a general framework for studying this complicated problem that decouples the product ranking problem for a single customer from the online matching of products to multiple customers over time. They also develop a better algorithm for the single-customer product ranking problem under well-studied cascade-click models. Finally, they introduce a model where the products are also arriving over time and cannot be included in the search rankings until they arrive.more » « less
-
Problem definition: We study scheduling multi-class impatient customers in parallel server queueing systems. At the time of arrival, customers are identified as being one of many classes, and the class represents the service and patience time distributions as well as cost characteristics. From the system’s perspective, customers of the same class at time of arrival get differentiated on their residual patience time as they wait in queue. We leverage this property and propose two novel and easy-to-implement multi-class scheduling policies. Academic/practical relevance: Scheduling multi-class impatient customers is an important and challenging topic, especially when customers’ patience times are nonexponential. In these contexts, even for customers of the same class, processing them under the first-come, first-served (FCFS) policy is suboptimal. This is because, at time of arrival, the system only knows the overall patience distribution from which a customer’s patience value is drawn, and as time elapses, the estimate of the customer’s residual patience time can be further updated. For nonexponential patience distributions, such an update indeed reveals additional information, and using this information to implement within-class prioritization can lead to additional benefits relative to the FCFS policy. Methodology: We use fluid approximations to analyze the multi-class scheduling problem with ideas borrowed from convex optimization. These approximations are known to perform well for large systems, and we use simulations to validate our proposed policies for small systems. Results: We propose a multi-class time-in-queue policy that prioritizes both across customer classes and within each class using a simple rule and further show that most of the gains of such a policy can be achieved by deviating from within-class FCFS for at most one customer class. In addition, for systems with exponential patience times, our policy reduces to a simple priority-based policy, which we prove is asymptotically optimal for Markovian systems with an optimality gap that does not grow with system scale. Managerial implications: Our work provides managers ways of improving quality of service to manage parallel server queueing systems. We propose easy-to-implement policies that perform well relative to reasonable benchmarks. Our work also adds to the academic literature on multi-class queueing systems by demonstrating the joint benefits of cross- and within-class prioritization. Funding: A. Bassamboo received financial support from the National Science Foundation [Grant CMMI 2006350]. C. (A.) Wu received financial support from the Hong Kong General Research Fund [Early Career Scheme, Project 26206419]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2023.1190 .more » « less
-
Motivated by applications from gig economy and online marketplaces, we study a two-sided queueing system under joint pricing and matching controls. The queueing system is modeled by a bipartite graph, where the vertices represent customer or server types and the edges represent compatible customer-server pairs. We propose a threshold-based two-price policy and queue length-based maximum-weight matching policy and show that it achieves a near-optimal profit. We study the system under the large-scale regime, wherein the arrival rates are scaled up, and under the large-market regime, wherein both the arrival rates and numbers of customer and server types increase. We show that two-price policy is a primary driver for optimality in the large-scale regime. We demonstrate the advantage of maximum-weight matching with respect to the number of customer and server types. Concurrently, we show that the interplay of pricing and matching is crucial for optimality in the large-market regime.more » « less
-
null (Ed.)We consider the problem faced by a service platform that needs to match limited supply with demand while learning the attributes of new users to match them better in the future. We introduce a benchmark model with heterogeneous workers (demand) and a limited supply of jobs that arrive over time. Job types are known to the platform, but worker types are unknown and must be learned by observing match outcomes. Workers depart after performing a certain number of jobs. The expected payoff from a match depends on the pair of types, and the goal is to maximize the steady-state rate of accumulation of payoff. Although we use terminology inspired by labor markets, our framework applies more broadly to platforms where a limited supply of heterogeneous products is matched to users over time. Our main contribution is a complete characterization of the structure of the optimal policy in the limit that each worker performs many jobs. The platform faces a tradeoff for each worker between myopically maximizing payoffs (exploitation) and learning the type of the worker (exploration). This creates a multitude of multiarmed bandit problems, one for each worker, coupled together by the constraint on availability of jobs of different types (capacity constraints). We find that the platform should estimate a shadow price for each job type and use the payoffs adjusted by these prices first to determine its learning goals and then for each worker (i) to balance learning with payoffs during the exploration phase and (ii) to myopically match after it has achieved its learning goals during the exploitation phase.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1007/s11134-023-09892-w
