AN IMPROVED HEURISTIC FOR STAFFING TELEPHONE CALL CENTERS WITH LIMITED OPERATING HOURS

Many telephone call centers that experience cyclic and random customer demand adjust their staffing over the day in an attempt to provide a consistent target level of customer service. The standard and widely used staffing method, which we call the stationary independent period by period (SIPP) approach, divides the workday into planning periods and uses a series of stationary independent Erlang‐c queuing models—one for each planning period—to estimate minimum staffing needs. Our research evaluates and improves upon this commonly used heuristic for those telephone call centers with limited hours of operation during the workday. We show that the SIPP approach often suggests staffing that is substantially too low to achieve the targeted customer service levels (probability of customer delay) during critical periods. The major reasons for SIPP‘ s shortfall are as follows: (1) SIPP's failure to account for the time lag between the peak in customer demand and when system congestion actually peaks; and (2) SIPP’ s use of the planning period average arrival rate, thereby assuming that the arrival rate is constant during the period. We identify specific domains for which SIPP tends to suggest inadequate staffing. Based on an analysis of the factors that influence the magnitude of the lag in infinite server systems that start empty and idle, we propose and test two simple “lagged” SIPP modifications that, in most situations, consistently achieve the service target with only modest increases in staffing.     

A Queueing Theoretic Approach to Set Staffing Levels in Time‐Dependent Dual‐Class Service Systems*

This article addresses the optimal staffing problem for a nonpreemptive priority queue with two customer classes and a time‐dependent arrival rate. The problem is related to several important service settings such as call centers and emergency departments where the customers are grouped into two classes of “high priority” and “low priority,” and the services are typically evaluated according to the proportion of customers who are responded to within targeted response times. To date, only approximation methods have been explored to generate staffing requirements for time‐dependent dual‐class services, but we propose a tractable numerical approach to evaluate system behavior and generate safe minimum staffing levels using mixed discrete‐continuous time Markov chains (MDCTMCs). Our approach is delicate in that it accounts for the behavior of the system under a number of different rules that may be imposed on staff if they are busy when due to leave and involves explicitly calculating delay distributions for two customer classes. Ultimately, we embed our methodology in a proposed extension of the Euler method, coined Euler Pri, that can cope with two customer classes, and use it to recommend staffing levels for the Welsh Ambulance Service Trust (WAST).     

Real‐Time Delay Estimation Based on Delay History in Many‐Server Service Systems with Time‐Varying Arrivals

Motivated by interest in making delay announcements in service systems, we study real‐time delay estimators in many‐server service systems, both with and without customer abandonment. Our main contribution here is to consider the realistic feature of time‐varying arrival rates. We focus especially on delay estimators exploiting recent customer delay history. We show that time‐varying arrival rates can introduce significant estimation bias in delay‐history‐based delay estimators when the system experiences alternating periods of overload and underload. We then introduce refined delay‐history estimators that effectively cope with time‐varying arrival rates together with non‐exponential service‐time and abandonment‐time distributions, which are often observed in practice. We use computer simulation to verify that our proposed estimators outperform several natural alternatives.     

Practical scheduling for call center operations

A practical spreadsheet-based scheduling method is developed to determine the optimal allocation of service agents to candidate tour types and start times in an inbound call center. A stationary Markovian queueing model with customer abandonment is employed to determine required staffing levels for a sequence of time intervals with varying call volumes, handling times, and relative agent availabilities. These staffing requirements populate a quadratic programming model for determining the distribution of agent tours that will maximize the fraction of offered calls beginning service within a target response time, subject to side constraints on tour type quantities. The optimal distribution is scaled to reflect the total number of scheduled agents, and a near-optimal integer solution is derived using rounding thresholds found by successive one-dimensional searches. This novel approach has been successfully implemented in large service centers at Qwest Communications and could easily be adapted to other operational environments.     

Flexible Servers in Understaffed Tandem Lines

We study the dynamic assignment of cross‐trained servers to stations in understaffed lines with finite buffers. Our objective is to maximize the production rate. We identify optimal server assignment policies for systems with three stations, two servers, different flexibility structures, and either deterministic service times and arbitrary buffers or exponential service times and small buffers. We use these policies to develop server assignment heuristics for Markovian systems with larger buffer sizes that appear to yield near‐optimal throughput. In the deterministic setting, we prove that the best possible production rate with full server flexibility and infinite buffers can be attained with partial flexibility and zero buffers, and we identify the critical skills required to achieve this goal. We then present numerical results showing that these critical skills, employed with an effective server assignment policy, also yield near‐optimal throughput in the Markovian setting, even for small buffer sizes. Thus, our results suggest that partial flexibility is sufficient for near‐optimal performance, and that flexibility structures that are effective for deterministic and infinite‐buffered systems are also likely to perform well for finite‐buffered stochastic systems.     

Hierarchical Multi‐skill Resource Assignment in the Telecommunications Industry

We formulate a discrete time Markov decision process for a resource assignment problem for multi‐skilled resources with a hierarchical skill structure to minimize the average penalty and waiting costs for jobs with different waiting costs and uncertain service times. In contrast to most queueing models, our application leads to service times that are known before the job is actually served but only after it is accepted and assigned to a server. We formulate the corresponding Markov decision process, which is intractable for problems of realistic size due to the curse of dimensionality. Using an affine approximation of the bias function, we develop a simple linear program that yields a lower bound for the minimum average costs. We suggest how the solution of the linear program can be used in a simple heuristic and illustrate its performance in numerical examples and a case study.     

Modeling and Managing the Percentage of Satisfied Customers in Hidden and Revealed Waiting Line Systems

We perform an analysis of various queueing systems with an emphasis on estimating a single performance metric. This metric is defined to be the percentage of customers whose actual waiting time was less than their individual waiting time threshold. We label this metric the Percentage of Satisfied Customers (PSC.) This threshold is a reflection of the customers' expectation of a reasonable waiting time in the system given its current state. Cases in which no system state information is available to the customer are referred to as “hidden queues.” For such systems, the waiting time threshold is independent of the length of the waiting line, and it is randomly drawn from a distribution of threshold values for the customer population. The literature generally assumes that such thresholds are exponentially distributed. For these cases, we derive closed form expressions for our performance metric for a variety of possible service time distributions. We also relax this assumption for cases where service times are exponential and derive closed form results for a large class of threshold distributions. We analyze such queues for both single and multi‐server systems. We refer to cases in which customers may observe the length of the line as “revealed” queues.“ We perform a parallel analysis for both single and multi‐server revealed queues. The chief distinction is that for these cases, customers may develop threshold values that are dependent upon the number of customers in the system upon their arrival. The new perspective this paper brings to the modeling of the performance of waiting line systems allows us to rethink and suggest ways to enhance the effectiveness of various managerial options for improving the service quality and customer satisfaction of waiting line systems. We conclude with many useful insights on ways to improve customer satisfaction in waiting line situations that follow directly from our analysis.     

Schedule Recovery: Unplanned Absences in Service Operations*

The U.S. service sector loses 2.3% of all scheduled labor hours to unplanned absences, but in some industries, the total cost of unplanned absences approaches 20% of payroll expense. The principal reasons for unscheduled absences (personal illness and family issues) are unlikely to abate anytime soon. Despite this, most labor scheduling systems continue to assume perfect attendance. This oversight masks an important but rarely addressed issue in services management: how to recover from short‐notice, short‐term reductions in planned capacity. In this article, we model optimal responses to unplanned employee absences in multi‐server queueing systems that provide discrete, pay‐per‐use services for impatient customers. Our goal is to assess the performance of alternate absence recovery strategies under various staffing and scheduling regimes. We accomplish this by first developing optimal labor schedules for hypothetical service environments with unreliable workers. We then simulate unplanned employee absences, apply an absence recovery model, and compute system profits. Our absence recovery model utilizes recovery strategies such as holdover overtime, call‐ins, and temporary workers. We find that holdover overtime is an effective absence recovery strategy provided sufficient reserve capacity (maximum allowable work hours minus scheduled hours) exists. Otherwise, less precise and more costly absence recovery methods such as call‐ins and temporary help service workers may be needed. We also find that choices for initial staffing and scheduling policies, such as planned overtime and absence anticipation, significantly influence the likelihood of successful absence recovery. To predict the effectiveness of absence recovery policies under alternate staffing/scheduling strategies and operating environments, we propose an index based on initial capacity reserves.     

Cross‐training Decisions in Field Services with Three Job Types and Server–Job Mismatch*

To be cost‐effective, field service managers must balance the high cost of machine downtime with the high cost of cross‐training technicians in multiple skills. We study a field service system with three job types requiring three different skills. Each server has a primary skill, the cost of which is considered sunk, and up to two secondary skills, which is a managerial decision. We model two important characteristics that distinguish field services: server–job mismatch and the ratio of travel time to service time. We use a queueing framework and simulation to study three cross‐training decisions: the number of servers cross‐trained in secondary skills, the number of secondary skills each server should have, and the efficiency in each secondary skill. We find that complete cross‐training is cost‐effective in some field service situations. Typically, efficiency in secondary skills must be close to 100%, but when the probability of mismatch is high and the ratio of travel time to service time is high, efficiency in secondary skills must be less than 100%.     

Pareto‐Improving Contracts for Express Package Delivery Services

We address the problem of an express package delivery company in structuring a long‐term customer contract whose terms may include prices that differ by day‐of‐week and by speed‐of‐service. The company traditionally offered speed‐of‐service pricing to its customers, but without day‐of‐week differentiation, resulting in customer demands with considerable day‐of‐week seasonality. The package delivery company hoped that using day‐of‐week and speed‐of‐service price differentiation for contract customers would induce these customers to adjust their demands to become counter‐cyclical to the non‐contract demand. Although this usually cannot be achieved by pricing alone, we devise an approach that utilizes day‐of‐week and speed‐of‐service pricing as an element of a Pareto‐improving contract. The contract provides the lowest‐cost arrangement for the package delivery company while ensuring that the customer is at least as well off as he would have been under the existing pricing structure. The contract pricing smoothes the package delivery company's demand and reduces peak requirements for transport capacity. The latter helps to decrease capital costs, which may allow a further price reduction for the customer. We formulate the pricing problem as a biconvex optimization model, and present a methodology for designing the contract and numerical examples that illustrate the achievable savings.     

Scheduling of Multi‐skilled Staff Across Multiple Locations

We address the problem of assigning airline customer service agents (CSAs) to tasks related to departing flights, such as selling tickets and collecting boarding cards, at an international terminal of a large airport. The airline specifies minimum and target levels of staff and required (or desired) types and levels of skills for each location in each time period. The assignment problem is complicated by staff heterogeneity, time required for moves between locations, and lunch and rest‐break requirements. We present a mixed‐integer formulation that considers both staffing shortages and skills mismatches and show that the problem is NP‐hard. We derive valid inequalities that tighten the bounds within a branch‐and‐cut procedure, enabling us to obtain near‐optimal solutions for problems of realistic size very quickly. We also present a generalization to simultaneously optimize shift starting times and task assignments, which can aid in longer term workforce planning. Finally, we utilize our procedure to obtain managerial insights regarding the benefits of flexibility derived from more highly skilled staff, allowing more frequent moves, and choices of shift starting times. We also demonstrate the benefits of our procedure vs. a heuristic that mimics what an experienced scheduler might choose.     

Subsampling Intervals in Autoregressive Models with Linear Time Trend

A new method is proposed for constructing confidence intervals in autoregressive models with linear time trend. Interest focuses on the sum of the autoregressive coefficients because this parameter provides a useful scalar measure of the long‐run persistence properties of an economic time series. Since the type of the limiting distribution of the corresponding OLS estimator, as well as the rate of its convergence, depend in a discontinuous fashion upon whether the true parameter is less than one or equal to one (that is, trend‐stationary case or unit root case), the construction of confidence intervals is notoriously difficult. The crux of our method is to recompute the OLS estimator on smaller blocks of the observed data, according to the general subsampling idea of Politis and Romano (1994a), although some extensions of the standard theory are needed. The method is more general than previous approaches in that it works for arbitrary parameter values, but also because it allows the innovations to be a martingale difference sequence rather than i.i.d. Some simulation studies examine the finite sample performance.     

Improving Voting Systems through Service‐Operations Management

We apply service‐operations‐management concepts to improve the efficiency and equity of voting systems. Recent elections in the United States and elsewhere have been plagued by long lines, excessive waiting times, and perceptions of unfairness. We build models for the waiting lines at voting precincts using both traditional steady‐state queueing methods and simulation models. We develop solution methods to allocate voting machines optimally to precincts. Our objective functions consider both the efficiency and the equity of the voting system. We compare our allocation algorithm to several competing methods, including those used in practice. We examine several different strategies for improving voting operations on both the demand and the capacity side of voting systems, and we present a complete case study of applying our method to data from the 2008 election for Franklin County, Ohio. We conclude that our method is superior to existing polices in terms of efficiency and equity and that it is robust in terms of uncertainties regarding turnout rates on Election Day. We also suggest several operational improvements to the voting process drawn from the service‐operations literature.     

Intelligent Procedures for Intra‐Day Updating of Call Center Agent Schedules

For nearly all call centers, agent schedules are typically created several days or weeks before the time that agents report to work. After schedules are created, call center resource managers receive additional information that can affect forecasted workload and resource availability. In particular, there is significant evidence, both among practitioners and in the research literature, suggesting that actual call arrival volumes early in a scheduling period (typically an individual day or week) can provide valuable information about the call arrival pattern later in the same scheduling period. In this paper, we develop a flexible and powerful heuristic framework for managers to make intra‐day resource adjustment decisions that take into account updated call forecasts, updated agent requirements, existing agent schedules, agents' schedule flexibility, and associated incremental labor costs. We demonstrate the value of this methodology in managing the trade‐off between labor costs and service levels to best meet variable rates of demand for service, using data from an actual call center.     

APPROXIMATIONS FOR THE GI/G/m QUEUE

Queueing models can usefully represent production systems experiencing congestion due to irregular flows, but exact analyses of these queueing models can be difficult. Thus it is natural to seek relatively simple approximations that are suitably accurate for engineering purposes. Here approximations for a basic queueing model are developed and evaluated. The model is the GI/G/m queue, which has m identical servers in parallel, unlimited waiting room, and the first-come first-served queue discipline, with service and interarrival times coming from independent sequences of independent and identically distributed random variables with general distributions. The approximations depend on the general interarrival-time and service-time distributions only through their first two moments. The main focus is on the expected waiting time and the probability of having to wait before beginning service, but approximations are also developed for other congestion measures, including the entire distributions of waiting time, queue-length and number in system. These relatively simple approximations are useful supplements to algorithms for computing the exact values that have been developed in recent years. The simple approximations can serve as starting points for developing approximations for more complicated systems for which exact solutions are not yet available. These approximations are especially useful for incorporating GI/G/m models in larger models, such as queueing networks, wherein the approximations can be components of rapid modeling tools.     

Panel Size and Overbooking Decisions for Appointment‐Based Services under Patient No‐Shows

Many service systems that work with appointments, particularly those in healthcare, suffer from high no‐show rates. While there are many reasons why patients become no‐shows, empirical studies found that the probability of a patient being a no‐show typically increases with the patient's appointment delay, i.e., the time between the call for the appointment and the appointment date. This paper investigates how demand and capacity control decisions should be made while taking this relationship into account. We use stylized single server queueing models to model the appointments scheduled for a provider, and consider two different problems. In the first problem, the service capacity is fixed and the decision variable is the panel size; in the second problem, both the panel size and the service capacity (i.e., overbooking level) are decision variables. The objective in both cases is to maximize some net reward function, which reduces to system throughput for the first problem. We give partial or complete characterizations for the optimal decisions, and use these characterizations to provide insights into how optimal decisions depend on patient's no‐show behavior in regards to their appointment delay. These insights especially provide guidance to service providers who are already engaged in or considering interventions such as sending reminders in order to decrease no‐show probabilities. We find that in addition to the magnitudes of patient show‐up probabilities, patients' sensitivity to incremental delays is an important determinant of how demand and capacity decisions should be adjusted in response to anticipated changes in patients' no‐show behavior.     

基于“云排队”的经济效益分析

随着各种云智能排队预约系统的推出，顾客“排队难”的问题得到了极大的缓解。其简便的操作方式，受到了大量商家与顾客的好评。本文基于当下热门服务行业的运营模式，建立排队博弈模型，从理论上分析了“云排队”平台对顾客以及服务商的影响。本文主要得到了以下几个结果：(1) 给出了两类顾客的纳什均衡策略(包括进队策略与最优到达间隔)；(2) 发现当前的预约机制能够提高“线上”顾客的效用，并刺激更多的线上顾客购买该产品，但同时也降低了“线下”顾客的进队意愿；(3) 发现在云排队机制下服务商的收益将大大提高，同时该机制将刺激服务商收取更高的服务费用。     

Nonparametric Test for Causality with Long‐range Dependence

This paper introduces a nonparametric Granger‐causality test for covariance stationary linear processes under, possibly, the presence of long‐range dependence. We show that the test is consistent and has power against contiguous alternatives converging to the parametric rate T^−1/2. Since the test is based on estimates of the parameters of the representation of a VAR model as a, possibly, two‐sided infinite distributed lag model, we first show that a modification of Hannan's (1963, 1967) estimator is root‐ T consistent and asymptotically normal for the coefficients of such a representation. When the data are long‐range dependent, this method of estimation becomes more attractive than least squares, since the latter can be neither root‐ T consistent nor asymptotically normal as is the case with short‐range dependent data.     

Experimental Results Indicating Lattice‐Dependent Policies May Be Optimal for General Assemble‐To‐Order Systems

We consider an assemble‐to‐order (ATO) system with multiple products, multiple components which may be demanded in different quantities by different products, possible batch ordering of components, random lead times, and lost sales. We model the system as an infinite‐horizon Markov decision process under the average cost criterion. A control policy specifies when a batch of components should be produced, and whether an arriving demand for each product should be satisfied. Previous work has shown that a lattice‐dependent base‐stock and lattice‐dependent rationing (LBLR) policy is an optimal stationary policy for a special case of the ATO model presented here (the generalized M‐system). In this study, we conduct numerical experiments to evaluate the use of an LBLR policy for our general ATO model as a heuristic, comparing it to two other heuristics from the literature: a state‐dependent base‐stock and state‐dependent rationing (SBSR) policy, and a fixed base‐stock and fixed rationing (FBFR) policy. Remarkably, LBLR yields the globally optimal cost in each of more than 22,500 instances of the general problem, outperforming SBSR and FBFR with respect to both objective value (by up to 2.6% and 4.8%, respectively) and computation time (by up to three orders and one order of magnitude, respectively) in 350 of these instances (those on which we compare the heuristics). LBLR and SBSR perform significantly better than FBFR when replenishment batch sizes imperfectly match the component requirements of the most valuable or most highly demanded product. In addition, LBLR substantially outperforms SBSR if it is crucial to hold a significant amount of inventory that must be rationed.     

A Model for Planning Resource Requirements in Health Care Organizations

In this paper we present a general model and solution methodology for planning resource requirements (i.e., capacity) in health care organizations. To illustrate the general model, we consider two specific applications: a blood bank and a health maintenance organization (HMO). The blood bank capacity planning problem involves determining the number of donor beds required and determining the size of the nursing and support staff necessary. Capacity must be sufficient to handle the expected number of blood donors without causing excessive donor waiting times. Similar staff, equipment, and service level decisions arise in the HMO capacity planning problem. To determine resource requirements, we develop an optimization/queueing network model that minimizes capacity costs while controlling customer service by enforcing a set of performance constraints, such as setting an upper limit on the expected time a patient spends in the system. The queueing network model allows us to capture the stochastic behavior of health care systems and to measure customer service levels within the optimization framework.