On the single item fill rate for a finite horizon

We investigate expressions for expected item fill rate in a periodic inventory system. The typical treatment of fill rate found in many operations management texts assumes infinite horizon, independent and stationary demand. For the case when the horizon is finite, we show that the expected value of the actual fill rate is greater than the value given by the infinite horizon expression. The implication of our results is that an inventory manager in a finite horizon situation who uses the infinite horizon expression to set stocking levels will achieve a higher than desired expected fill rate at greater than necessary inventory expense.     

Probabilistic modeling of multiperiod service levels

This study investigates multiperiod service level (MSL) policies in supply chains facing a stochastic customer demand. The objective of the supply chains is to construct integrated replenishment plans that satisfy strict stockout-oriented performance measures which apply across a multiperiod planning horizon. We formulate the stochastic service level constraints for the fill rate, ready rate, and conditional expected stockout MSL policies. The modeling approach is based on the concept of service level trajectory and provides reformulations of the stochastic planning problems associated with each MSL policy that can be efficiently solved with off-the-shelf optimization solvers. The approach enables the handling of correlated and non-stationary random variables, and is flexible enough to accommodate the implementation of fair service level policies, the assignment of differentiated priority levels per products, or the introduction of response time requirements. We use an earthquake disaster management case study to show the applicability of the approach and derive practical implications about service level policies.     

Queues with waiting time dependent service

Motivated by service levels in terms of the waiting-time distribution seen, for instance, in call centers, we consider two models for systems with a service discipline that depends on the waiting time. The first model deals with a single server that continuously adapts its service rate based on the waiting time of the first customer in line. In the second model, one queue is served by a primary server which is supplemented by a secondary server when the waiting of the first customer in line exceeds a threshold. Using level crossings for the waiting-time process of the first customer in line, we derive steady-state waiting-time distributions for both models. The results are illustrated with numerical examples.     

Reservation and allocation policies for influenza vaccines

This research investigates the impact of alternative allocation mechanisms that can be employed in the context of vaccine inventory rationing. Available vaccine inventory can be allocated to arrivals from high priority (target groups such as healthcare professionals) and low priority (non-target groups) demand classes using Partitioned Allocation (PA), Standard Nesting (SN), and Theft Nesting (TN). In any one of the mechanisms, a part of the available inventory is reserved for the exclusive use of the high priority demand class. They differ, however, in how the unreserved portion of the inventory is utilized: Under PA, demand from the high (low) priority class consumes only the reserved (unreserved) quantity. Under SN, demand from the high priority class first consumes the reserved quantity; once and if this quantity is exhausted, high priority demand competes with low priority demand for the remaining inventory. Under TN the sequence of allocation is reversed: both demand classes first compete for the unreserved inventory. Once this portion of inventory is exhausted, high priority demand is fulfilled from the reserved inventory and low priority demand is rejected. We develop service level (probability of fulfilling the entire demand) and fill rate (fraction of demand fulfilled) expressions for all three allocation mechanisms. Based on these expressions, numerical analyses are conducted to illustrate which allocation mechanism a health planner should choose depending on the availability of vaccines, and how the health planner should set the reserved quantity for the high priority class. We observe that (1) there exist certain conditions under which one of the allocation mechanisms outperforms the others and (2) this effect is determined by the decision maker’s choice of the performance measure.     

Tail probabilities of low-priority waiting times and queue lengths in MAP/GI/1 queues

We consider the problem of estimating tail probabilities of waiting times in statistical multiplexing systems with two classes of sources – one with high priority and the other with low priority. The priority discipline is assumed to be nonpreemptive. Exact expressions for the transforms of these quantities are derived assuming that packet or cell streams are generated by Markovian Arrival Processes (MAPs). Then a numerical investigation of the large-buffer asymptotic behavior of the the waiting-time distribution for low-priority sources shows that these asymptotics are often non-exponential.     

Asymptotics for M/G/1 low-priority waiting-time tail probabilities

We consider the classical M/G/1 queue with two priority classes and the nonpreemptive and preemptive-resume disciplines. We show that the low-priority steady-state waiting-time can be expressed as a geometric random sum of i.i.d. random variables, just like the M/G/1 FIFO waiting-time distribution. We exploit this structures to determine the asymptotic behavior of the tail probabilities. Unlike the FIFO case, there is routinely a region of the parameters such that the tail probabilities have non-exponential asymptotics. This phenomenon even occurs when both service-time distributions are exponential. When non-exponential asymptotics holds, the asymptotic form tends to be determined by the non-exponential asymptotics for the high-priority busy-period distribution. We obtain asymptotic expansions for the low-priority waiting-time distribution by obtaining an asymptotic expansion for the busy-period transform from Kendall's functional equation. We identify the boundary between the exponential and non-exponential asymptotic regions. For the special cases of an exponential high-priority service-time distribution and of common general service-time distributions, we obtain convenient explicit forms for the low-priority waiting-time transform. We also establish asymptotic results for cases with long-tail service-time distributions. As with FIFO, the exponential asymptotics tend to provide excellent approximations, while the non-exponential asymptotics do not, but the asymptotic relations indicate the general form. In all cases, exact results can be obtained by numerically inverting the waiting-time transform. This revised version was published online in June 2006 with corrections to the Cover Date.     

A partial-postponement decision cost model

A number of factors, including product proliferation and increased customer service-level requirements, have led many companies to consider adopting postponement as a supply chain strategy. Packaging postponement is the process of delaying packaging of a common item into a final product configuration until the customer order is received. For a given product, a portion of demand is known with a high level of certainty and would not benefit from postponement. The remaining portion of demand is known with little certainty and would benefit from delaying the differentiating stage of the operation until demand is known. We develop a single-period, two-product, order-up-to cost model to aid in setting the levels of finished-goods inventory and postponement capacity. Minimum-cost optimal solutions to inventory levels and capacity are obtained by solving the derived analytical expressions using a non-linear programming formulation. We examine the sensitivity of the model to different levels of the model parameters to generate managerial insights beyond those of previous work. We show that changing product value, packaging cost, cost of postponement, holding cost, fill rate, and demand correlation can decrease expected total cost and increase postponement capacity.     

Nonpreemptive Multi-Server Priority Queues

Hokstad recently published an approximate method for calculating the behaviour of an M/G/m queue. This note applies his results to the nonpreemptive priority situation with two priority classes having the same service-time distribution. Laplace transforms and the first two moments of the waiting-time distributions are given.     

Using different response-time requirements to smooth time-varying demand for service

Many service systems have demand that varies significantly by time of day, making it costly to provide sufficient capacity to be able to respond very quickly to each service request. Fortunately, however, different service requests often have very different response-time requirements. Some service requests may need immediate response, while others can tolerate substantial delays. Thus it is often possible to smooth demand by partitioning the service requests into separate priority classes according to their response-time requirements. Classes with more stringent performance requirements are given higher priority for service. Lower capacity may be required if lower-priority-class demand can be met during off-peak periods. We show how the priority classes can be defined and the resulting required fixed capacity can be determined, directly accounting for the time-dependent behavior. For this purpose, we exploit relatively simple analytical models, in particular, M_t/G/∞ and deterministic offered-load models. The analysis also provides an estimate of the capacity savings that can be obtained from partitioning time-varying demand into priority classes.     

The impact of data collection on fill rate performance in the (R,s, Q) inventory model

In this paper we consider the determination of the reorder point s in an (R, s, Q) inventory model subject to a fill rate service level constraint. We assume that the underlying demand process is a compound renewal process. We then derive an approximation method to compute the reorder level such that a target service level is achieved. Restrictions on the input parameters are given, within which this method is applicable. Moreover, we will investigate the effects on the fill rate performance in case the underlying demand process is indeed a compound renewal process, while the demand process is modelled as a discrete-time demand process. That is, the time axis is divided in time units (for example, days) and demands per time unit are independent and identically distributed random variables. It will be shown that smooth and erratic behaviour of the inter-arrival times have different impacts on the performance of the fill rate when demand is modelled as a discrete-time process and in case the underlying demand process is a compound renewal process.     

On the exact calculation of the fill rate in a periodic review inventory policy under discrete demand patterns

The primary goal of this paper is the development of a generalized method to compute the fill rate for any discrete demand distribution in a periodic review policy. The fill rate is defined as the fraction of demand that is satisfied directly from shelf. In the majority of related work, this service metric is computed by using what is known as the traditional approximation, which calculates the fill rate as the complement of the quotient between the expected unfulfilled demand and the expected demand per replenishment cycle, instead of focusing on the expected fraction of fulfilled demand. This paper shows the systematic underestimation of the fill rate when the traditional approximation is used, and revises both the foundations of the traditional approach and the definition of fill rate itself. As a result, this paper presents the following main contributions: (i) a new exact procedure to compute the traditional approximation for any discrete demand distribution; (ii) a more suitable definition of the fill rate in order to ignore those cycles without demand; and (iii) a new standard procedure to compute the fill rate that outperforms previous approaches, especially when the probability of zero demand is substantial. This paper focuses on the traditional periodic review, order up to level system under any uncorrelated, discrete and stationary demand pattern for the lost sales scenario.     

Supply chain inventory optimization with two customer classes in discrete time

In this paper we consider a single-item inventory system where two demand classes with different service requirements are satisfied from a common inventory. A critical level, reorder point, order quantity or (s, q, k) policy is in use. The time axis is divided into discrete time units, which is a common characteristic of many real-life supply-chain processes. The inventory process within the lead time of a replenishment order is modelled as a sequence of (1) an ordinary renewal process and (2) two alternating renewal processes. Approximations are developed for the demand class-specific fill rates and the probability distribution of the waiting time of low priority customer orders. This waiting time distribution is used for the inventory allocation in a two-stage supply chain.     

Waiting times in queueing networks with a single shared server

We study a queueing network with a single shared server that serves the queues in a cyclic order. External customers arrive at the queues according to independent Poisson processes. After completing service, a customer either leaves the system or is routed to another queue. This model is very generic and finds many applications in computer systems, communication networks, manufacturing systems, and robotics. Special cases of the introduced network include well-known polling models, tandem queues, systems with a waiting room, multi-stage models with parallel queues, and many others. A complicating factor of this model is that the internally rerouted customers do not arrive at the various queues according to a Poisson process, causing standard techniques to find waiting-time distributions to fail. In this paper, we develop a new method to obtain exact expressions for the Laplace–Stieltjes transforms of the steady-state waiting-time distributions. This method can be applied to a wide variety of models which lacked an analysis of the waiting-time distribution until now.     

Optimal operational service levels in vendor managed inventory contracts - an exact approach

Vendor Managed Inventory (VMI) contracts are anchored on a fill rate at which the vendor is expected to meet the end-customer demand. Violations of this contracted fill rate due to excess and insufficient inventory are both penalized, often in a linear, but asymmetric manner. To minimize these costs, the vendor needs to maintain an operational fill rate that is different from the contracted fill rate. We model, analyze and solve an optimization problem that determines this operational fill rate and the associated optimal inventory decision. We establish that, for some special, yet popular, models of demand (e.g. truncated normal, gamma, Weibull and uniform distributions), the optimal solution can be derived in closed form and computed precisely. For other demand distributions, either the optimization problem becomes ill-defined or we may need to use approximate solution methods. An extensive computational study reveals that, for realistic values of problem parameters, the operational fill rate is often larger (by as much as 20%) than the contracted service level, possibly explaining the inventory glut commonly observed in real-world VMI systems.     

Determining the fill rate for a periodic review inventory policy with capacitated replenishments,lost sales and zero lead time

In this paper we consider a periodic review order-up-to inventory system with capacitated replenishments, lost sales and zero lead time. We consider discrete demand. It is shown that the initial stock levels of the different review periods form a Markov chain and we determine the transition matrix. Furthermore we study for what probability mass functions of the review period demand the Markov chain has a unique stationary distribution. Finally, we present a method to determine the fill rate.     

Improving resource utilization for diagnostic services through flexible inpatient scheduling: A method for improving resource utilization

We present a method to increase the utilization of and reduce the waiting times for an under-capacitated diagnostic resource in the presence of uncertain demand with several priority levels. We consider the case of a computed tomography (CT) scanning department that services both high-priority in-patients and lower priority outpatients. Current practice calls for all in-patient demand to be met on the day of the request. Our proposal looks at the benefit of reserving space for carrying over a percentage of non-emergency in-patient demand to the next day and utilizing a pool of on-call outpatients who can respond quickly to available capacity. We formulate and solve an optimization problem that returns a reservation policy that minimizes unused capacity subject to an overtime constraint. We use a simulation to demonstrate a significant reduction in the growth rate of outpatient waiting time resulting from using the proposed method and investigate the sensitivity of results to several model assumptions.     

Queueing analysis on a single-station make-to-stock/make-to-order inventory-production system

We consider a one-station production system that produces standard products for ordinary demands and custom products for specific demands. In this system, the workstation has two manufacturing modes. In mode 1, it produces standard products and, in mode 2, it produces custom products by performing the additional alternating works on one existing finished standard product. Base-stock control policy is applied to control the production of standard products. The fill rate of the ordinary demand and the on-time-delivery-rate of the specific demand are considered as the measures of the qualities of service. By assuming an Markovian system, qualities of service under base-stock policy are obtained; furthermore, the optimal base-stock level can be obtained numerically under the requirements on the qualities of services.     

Bandwidth packing with priority classes

The bandwidth packing problem is defined as the selection and routing of messages from a given list of messages with prespecified requirements on demand for bandwidth. The messages have to be routed over a network with given topology so that the generated revenue is maximized. Messages to be routed are classified into two priority classes. An integer programming based formulation of this problem is proposed and a Lagrangean relaxation based methodology is described for solving this problem. A general purpose heuristic is then developed for generating feasible solutions of good quality. Several numerical experiments are conducted using a number of problem parameters such as number of messages, ratio of messages for lower and higher priority classes, capacity of links, and demand distribution of messages belonging to different classes and high quality solutions to the priority bandwidth packing problem are generated under the different situations.     

Large deviations for renewal processes

We investigate large deviations for the empirical measure of the forward and backward recurrence time processes associated with a classical renewal process with arbitrary waiting-time distribution. The Donsker-Varadhan theory cannot be applied in this case, and indeed it turns out that the large deviations rate functional differs from the one suggested by such a theory. In particular, a non-strictly convex and non-analytic rate functional is obtained.     

Stationary waiting time derivatives

We investigate the stability of waiting-time derivatives when inputs to a queueing system-service times and interarrival times-depend on a parameter. We give conditions under which the sequence of waiting-time derivatives admits a stationary distribution, and under which the derivatives converge to the stationary regime from all initial conditions. Further hypotheses ensure that the expectation of a stationary waiting-time derivative is, in fact, the derivative of the expected stationary waiting time. This validates the use of simulation-based infinitesimal perturbation analysis estimates with a variety of queueing processes.We examine waiting-time sequences satisfying recursive equations. Our basic assumption is that the input and its derivatives are stationary and ergodic. Under monotonicity conditions, the method of Loynes establishes the convergence of the derivatives. Even without such conditions, the derivatives obey a linear difference equation with random coefficients, and we exploit this fact to find stability conditions.