Search

We apply supervised learning to a general problem in queueing theory: using a neural net, we develop a fast and accurate predictor of the stationary system-length distribution of a GI/GI/1 queue—a fundamental queueing model for which no analytical solutions are available. To this end, we must overcome three main challenges: (i) generating a large library of training instances that cover a wide range of arbitrary interarrival and service time distributions, (ii) labeling the training instances, and (iii) providing continuous arrival and service distributions as inputs to the neural net. To overcome (i), we develop an algorithm to sample phase-type interarrival and service time distributions with complex transition structures. We demonstrate that our distribution-generating algorithm indeed covers a wide range of possible positive-valued distributions. For (ii), we label our training instances via quasi-birth-and-death(QBD) that was used to approximate PH/PH/1 (with phase-type arrival and service process) as labels for the training data. For (iii), we find that using only the first five moments of both the interarrival and service times distribution as inputs is sufficient to train the neural net. Our empirical results show that our neural model can estimate the stationary behavior of the GI/GI/1—far exceeding other available methods in terms of both accuracy and runtimes. History: Ram Ramesh, Area Editor for Data Science and Machine Learning. Funding: O. Baron received financial support from the Natural Sciences and Engineering Research Council of Canada (NERC) [Grant 458051]. D. Krass received financial support from the NERC [Grant 458395]. Supplemental Material: The software that supports the findings of this study is available within the paper and its Supplemental Information (https://pubsonline.informs.org/doi/suppl/10.1287/ijoc.2022.0263) as well as from the IJOC GitHub software repository (https://github.com/INFORMSJoC/2022.0263). The complete IJOC Software and Data Repository is available at https://informsjoc.github.io/. [ABSTRACT FROM AUTHOR]

This paper shows that operational flexibility interacting with informational uncertainty may lead to truthful information exchange in equilibrium even when the communication is nonbinding and unverifiable, i.e., 'cheap talk.' We consider a model consisting of a manufacturer releasing a new product with uncertain release date and demand, and a retailer who must determine the allocation of limited capacity between a preexisting third-party product and the manufacturer's new product that may or may not be released on time. The manufacturer has a private forecast about the likelihood of the product release and/or about the demand, which he shares (either truthfully or not) with the retailer. We show that under the 'traditional' supply chain structure (one-time opportunity to order) no truthful equilibrium can emerge. However, if (1) the supply chain structure allows for postponement, i.e., the ability to delay orders at a certain cost by the retailer, and (2) the manufacturer has informational uncertainty about the retailer's capacity, then truthful information exchange may emerge in equilibrium, where the manufacturer transmits his true forecast and the retailer treats the transmission as truthful. The genesis of this effect is preference reversal, where the manufacturer is not sure which way to distort the forecast to best motivate the retailer to wait for the new product. Thus, we show that a truth-revealing mechanism can emerge from a relatively rich setup featuring two-sided information asymmetry interacting with postponement. History: Accepted by Gad Allon, operations management. Keywords: forecast information sharing * cheap talk * supply chain flexibility * postponement * procurement timing, 1. Introduction Coordination between different levels of a supply chain is critical for the chain to operate effectively (see, e.g., Byrnes and Shapiro 1991, Kurt Salmon Associates 1993). One of [...]

This paper, motivated by a collaboration with a healthcare service provider, focuses on stochastic open-shop service networks with two objectives: more traditional macrolevel measures (such as minimizing total system time or minimizing total number of tardy customers) and the atypical microlevel measure of reducing the incidents of excessively long waits at any workstation within the process. While work-conserving policies are optimal for macrolevel measures, scheduling policies with strategic idleness (SI) might be helpful for microlevel measures. Using the empirical data obtained from the service provider, we provide statistical evidence that SI is used by its schedulers to manage the macro- and microlevel measures. However, the company has no specific rules on implementing SI and the schedulers make decisions based on their own experience. Our primary goal is to develop a systematic framework for the joint usage of SI with dynamic scheduling policies (DSPs). We suggest to use threshold-based policies to intelligently combine SI and DSPs and show that the resulting policies provide an efficient way to simultaneously address both macro- and microlevel measures. We build two simulation models: one based on empirical data and one based on a randomly generated open-shop network. We use both models to demonstrate that an open-shop service network can be systematically and effectively managed to deliver improved service level by using SI. Funding: The first three authors acknowledge financial support from their respective Natural Sciences and Engineering Research Council of Canada grants. Supplemental Material: The online appendix is abailable at https://doi.org/10.1287/msom.2016.0591. Keywords: strategic idleness * dynamic scheduling policy * open shop * healthcare * service network, 1. Introduction In this paper, we focus on stochastic open-shop processes operating under multiple objectives. The system objectives include a combination of the more traditional macrolevel measures (such as minimizing [...]

A cooperative-covering family of location problems is proposed in this paper. Each facility emits a (possibly non-physical) 'signal' which decays over the distance and each demand point observes the aggregate signal emitted by all facilities. It is assumed that a demand point is covered if its aggregate signal exceeds a given threshold; thus facilities cooperate to provide coverage, as opposed to the classical coverage location model where coverage is only provided by the closest facility. It is shown that this cooperative assumption is appropriate in a variety of applications. Moreover, ignoring the cooperative behavior (i.e., assuming the traditional individual coverage framework) leads to solutions that are significantly worse than the optimal cooperative cover solutions; this is illustrated with a case study of locating warning sirens in North Orange County, California. The problems are formulated, analyzed and solved in the plane for the Euclidean distance case. Optimal and heuristic algorithms are proposed and extensive computational experiments are reported. Keywords: Cover, planar location, cooperative cover, 1. Introduction Location models with cover objective are one of the 'classical' types of facility location models. Under this objective, a demand point is considered 'covered' if there is an [...]

To authenticate to the full-text of this article, please visit this link: http://dx.doi.org/10.1111/j.1540-5915.2009.00253.x Byline: Oded Berman (1), Dmitry Krass (1), Mozart B. C. Menezes (2) Keywords: Discrete; Discrete Location; Facilities/Equipment Planning; and Reliability and p-Median Abstract: ABSTRACT In this article, we analyze a location model where facilities may be subject to disruptions. Customers do not have advance information about whether a given facility is operational or not, and thus may have to visit several facilities before finding an operational one. The objective is to locate a set of facilities to minimize the total expected cost of customer travel. We decompose the total cost into travel, reliability, and information components. This decomposition allows us to put a value on the advance information about the states of facilities and compare it to the reliability and travel cost components, which allows a decision maker to evaluate which part of the system would benefit the most from improvements. The structure of optimal solutions is analyzed, with two interesting effects identified: facility centralization and co-location; both effects appear to be stronger than in the complete information case, where the status of each facility is known in advance. Author Affiliation: (1)Rotman School of Management, University of Toronto, 105 St. George Street, Toronto, Ontario, Canada M5S 3E6, e-mail:berman@rotman.utoronto.ca,krass@rotman.utoronto.ca (2)MIT-Zaragoza International Logistics Program, Zaragoza Logistics Center, Avenida Gomez Laguna 25, 1 Planta, 50009 Zaragoza, Spain, e-mail:mmenezes@zlc.edu.es Article History: [Received: November 2006; Accepted: June 2009.] Article note: [dagger]Corresponding author.

A location problem with stochastic demand and congestion where mobile servers respond to service calls originating from nodes is considered. The problem is of the set-covering type: only servers within the coverage radius of the demand-generating node may respond to a call. The service level constraint requires that at least one server must be available to respond to an arriving call, with some prespecified probability. The objective is to minimize the total number of servers. It is shown that earlier models quite often overestimate servers' availability and thus may lead to infeasible solutions (i.e., solutions that fail to satisfy the service level constraint). System stability conditions and lower bounds on system availability are developed by analyzing the underlying partially accessible queueing system. These lead to the development of two new models for which feasibility is guaranteed. Simulation-based computational experiments show that the proposed models achieve feasibility without significantly increasing the total number of servers. [Supplementary materials are available for this article. Go to the publisher's online edition of IIE Transactions for the following free supplemental resource: Appendix of Tables of Computational Results for Section 7.] Keywords: Location, network, queueing, congestion, 1. Introduction The fundamental problem facing emergency system planners can be described as follows. Customers residing at the nodes of a transportation network generate stochastic streams of service calls. These [...]

We analyze the problem of optimal location of a set of facilities in the presence of stochastic demand and congestion. Customers travel to the closest facility to obtain service; the [...]

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.ejor.2006.10.023 Byline: Opher Baron (a), Oded Berman (a), Dmitry Krass (a), Qian Wang (b) Keywords: Location; Heuristics; Facilities planning and design Abstract: This paper considers the problem of locating M facilities on the unit square so as to minimize the maximal demand faced by each facility subject to closest assignments and coverage constraints. Focusing on uniform demand over the unit square, we develop upper and lower bounds on feasibility of the problem for a given number of facilities and coverage radius. Based on these bounds and numerical experiments we suggest a heuristic to solve the problem. Our computational results show that the heuristic is very efficient, as the average gap between its solutions and the lower bound is 4.34%. Author Affiliation: (a) Joseph L. Rotman School of Management, University of Toronto, 105 St. George Street, Toronto, Ont., Canada M5S 3E6 (b) Management School of Graduate University, Chinese Academy of Sciences, 80 Zhongguancun East Road, Haidian District, Beijing 100080, PR China Article History: Received 26 January 2006; Accepted 12 October 2006

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.ejor.2006.07.021 Byline: Robert Aboolian (a), Oded Berman (b), Dmitry Krass (b) Keywords: Competitive facility location models; Facility design problems; Non-linear integer programming; Greedy heuristics; Worst-case bounds; Approximation Abstract: We develop a spatial interaction model that seeks to simultaneously optimize location and design decisions for a set of new facilities. The facilities compete for customer demand with pre-existing competitive facilities and with each other. The customer demand is assumed to be elastic, expanding as the utility of the service offered by the facilities increases. Increases in the utility can be achieved by increasing the number of facilities, design improvements, or locating facilities closer to the customer. We show that our model is able to capture some of the principal trade-offs involved in facility location and design decisions, including demand cannibalization, market expansion, and design/location trade-offs. Managerial insights are obtained through sensitivity analysis of the model and through several illustrative examples. An efficient near-optimal solution approach, with adjustable error bound, is developed for the special case where only a finite number of design alternatives are available. Several heuristic approaches capable of handling large instances are also presented. Author Affiliation: (a) College of Business Administration, California State University San Marcos, San Marcos, CA 92096, USA (b) Rotman School of Management, University of Toronto, 105 St. George Street, Toronto, Ont., Canada M5S 3E6 Article History: Received 24 March 2006; Accepted 19 July 2006

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.ejor.2005.10.075 Byline: Robert Aboolian (a), Oded Berman (b), Dmitry Krass (b) Keywords: Location; Integer programming; Competitive facility location models; Non-linear Knapsack problem; Alpha-optimal solutions; Greedy heuristics; Worst-case bounds Abstract: We consider a spatial interaction model for locating a set of new facilities that compete for customer demand with each other, as well as with some pre-existing facilities to capture the 'market expansion' and the 'market cannibalization' effects. Customer demand is assumed to be a concave non-decreasing function of the total utility derived by each customer from the service offered by the facilities. The problem is formulated as a non-linear Knapsack problem, for which we develop a novel solution approach based on constructing an efficient piecewise linear approximation scheme for the objective function. This allows us to develop exact and [alpha]-optimal solution approaches capable of dealing with relatively large-scale instances of the model. We also develop a fast Heuristic Algorithm for which a tight worst-case error bound is established. Author Affiliation: (a) College of Business Administration, California State University San Marcos, San Marcos, CA 92096, USA (b) Rotman School of Management, University of Toronto, 105 St. George Street, Toronto, Ont., Canada M5S 3E6 Article History: Received 1 March 2004; Accepted 6 October 2005

We analyze the problem of locating a set of service facilities on a network when the demand for service is stochastic and congestion may arise at the facilities. We consider two potential sources of lost demand: (i) demand lost due to insufficient coverage; and (ii) demand lost due to congestion. Demand loss due to insufficient coverage arises when a facility is located too far away from customer locations. The amount of demand lost is modeled as an increasing function of the travel distance. The second source of lost demand arises when the queue at a facility becomes too long. It is modeled as the proportion of balking customers in a Markovian queue with a fixed buffer length. The objective is to find the minimum number of facilities, and their locations, so that the amount of demand lost from either source does not exceed certain pre-set levels. After formulating the model, we derive and investigate several different integer programming formulations, focusing in particular on alternative representations of closest assignment constraints. We also investigate a wide variety of heuristic approaches, ranging from simple greedy-type heuristics, to heuristics based on time-limited branch and bound, tabu search, and random adaptive search heuristics. The results of an extensive set of computational experiments are presented and discussed., 1. Introduction A primary objective in locating service facilities is to 'capture' as much of the potential customer demand as possible. Most companies have a series of decision rules (based [...]