Capacity planning in networks of queues with manufacturing applications

This paper addresses capacity planning in systems that can be modeled as a network of queues. More specifically, we present an optimization model and solution methods for the minimum cost selection of capacity at each node in the network such that a set of system performance constraints is satisfied. Capacity is controlled through the mean service rate at each node. To illustrate the approach and how queueing theory can be used to measure system performance, we discuss a manufacturing model that includes upper limits on product throughput times and work-in-process in the system. Methods for solving capacity planning problems with continuous and discrete capacity options are discussed. We focus primarily on the discrete case with a concave cost function, allowing fixed charges and costs exhibiting economies of scale with respect to capacity to be handled.     

Better Optimization of Nonlinear Uncertain Systems (BONUS): A New Algorithm for Stochastic Programming Using Reweighting through Kernel Density Estimation

Given a series-parallel queueing network topology with exponential servers of finite capacity, a systematic design methodology is presented that approximately solves the optimal routing and buffer space allocation problems within the network. The multi-objective stochastic nonlinear programming problem in integer variables is described and a two-stage iterative optimization procedure is presented which interconnects the routing and buffer space allocation problems. The algorithmic procedure couples the Expansion method, a decomposition method for computing performance measures in queueing networks with finite capacity, along with Powell's unconstrained optimization procedure which allocates the buffers and a multi-variable search procedure for determining the routing probabilities. The effectiveness and efficiency of the resulting two-stage design methodology is tested and evaluated in a series of experimental designs along with simulations of the network topologies.     

Rapid improvement of stochastic networks using two-moment approximations

This paper investigates a method for rapidly improving the performance of stochastic networks. The method uses queueing network techniques based on parametric decomposition and two-moment approximations. It employs some recently devised developments to the approximations, especially in the context of closed queueing networks with synchronization stations. We describe these developments and explain how the method uses them. It appears to offer a substantial advantage in speed, and reduces the dependence on repetitive simulation runs for identifying opportunities to improve network performance. We present an example from airfield operations, in which the method evaluates the tradeoffs with respect to several strategies for enhancing airfield performance, such as the introduction of concurrent operations, the reduction of service variability, and the increase in airfield capacity.     

On the Full Idle Time in Multiphase Queueing Systems

The modern queueing theory is a powerful tool for a quantitative and qualitative analysis of communication systems, computer networks, transportation systems, and many other technical systems. The paper is designated to the analysis of queueing systems arising in the network theory and communications theory (such as the so-called multiphase queueing systems, tandem queues, or series of queueing systems). We present heavy traffic limit theorems for the full idle time in multiphase queueing systems. We prove functional limit theorems for values of the full idle time of a queueing system, which is its important probability characteristic. __________ Translated from Lietuvos Matematikos Rinkinys, Vol. 45, No. 3, pp. 367–386, July–September, 2005.     

Deadlock properties of queueing networks with finite capacities and multiple routing chains

Blocking in queueing network models with finite capacities can lead to deadlock situations. In this paper, deadlock properties are investigated in queueing networks with multiple routing chains. The necessary and sufficient conditions for deadlockfree queueing networks with blocking are provided. An optimization algorithm is presented for finding deadlock-free capacity assignments with the least total capacity. The optimization algorithm maps the queueing network into a directed graph and obtains the deadlock freedom conditions from a specified subset of cycles in the directed graph. In certain network topologies, the number of deadlock freedom conditions can be large, thus, making any optimization computationally expensive. For a special class of topologies, so-calledtandem networks, it is shown that a minimal capacity assignment can be directly obtained without running an optimization algorithm. Here, the solution to the minimal capacity assignment takes advantage of the regular topology of tandem networks.This work was supported by the National Science Foundation under Grant No. CCR-90-11981.     

Topological network design of general,finite, multi-server queueing networks

The topological network design of general service, finite waiting room, multi-server queueing networks is a complex optimization problem. Series, merge, and split topologies are examined using an approximation method to estimate the performance of these queueing networks and an iterative search methodology to find the optimal buffer allocation within the network. The coefficient of variation is shown to be a significant factor in the buffer allocation for multiple servers in uniform and bottleneck server networks. Extensive computational results are included to illustrate the symmetries and asymmetries in the buffer patterns which emerge from the series, merge, and splitting topologies.     

具有相型服务分布的排队系统的性能灵敏度分析

具有非线数服务分布的排队网络已被广泛应用于许多领域,如通讯网络和管理系统。本文借助于无穷小说矩阵摄动方法,研究了M/PH/1排队系统的稳态性能灵敏度分析问题,给出了性能灵敏度公式,并表明了稳态性能灵敏度很容易通过系统势能进行计算。同时,给出一种计算势能及性能导数的算法。这个算法可直接用于系统的控制与优化,因为它基于分析系统的一条单一样本轨道。最后提供一个数值例子来表明这个算法的应用。     

Error analysis for regenerative queueing estimators with special reference to gradient estimators via likelihood ratio

The performance of telecommunications systems is typically estimated (either analytically or by simulation) via queueing theoretic models. The gradient of the expected performance with respect to the various parameters (such as arrival rate or service rate) is very important as it not only measures the sensitivity to change, but is also needed for the solution of optimization problems. While the estimator for the expected performance is the sample mean of the simulation experiment, there are several possibilities for the estimator of the gradient. They include the obvious finite difference approximation, but also other recently advocated techniques, such as estimators derived from likelihood ratio transformations or from infinitesimal perturbations. A major problem in deciding upon which estimator to use and in planning the length of the simulation has been the scarcity of analytical error calculations for estimators of queueing models. It is this question that we answer in this paper for the waiting time moments (of arbitrary order) of theM / G / 1 queue by using the queueing analysis technique developed by Shalmon. We present formulas for the error variance of the estimators of expectation and of its gradient as a function of the simulation length; at arbitrary traffic intensity the formulas are recursive, while the heavy traffic approximations are explicit and of very simple form. For the gradient of the mean waiting time with respect to the arrival (or service) rate, and at 1 percent relative precision, the heavy traffic formulas show that the likelihood ratio estimator for the gradient reduces the length of the simulation required by the finite difference estimator by about one order of magnitude; further increasing the relative precision by a factor increases the reduction advantage by the same factor. At any relative precision, it exceeds the length of the simulation required for the estimation of the mean with the same relative precision by about one order of magnitude. While strictly true for theM / G / 1 queue, the formulas can also be used as guidelines in the planning of queueing simulations and of stochastic optimizations of complex queueing systems, particularly those with Poisson arivals.     

Performance optimization of queueing systems with perturbation realization

After the intensive studies of queueing theory in the past decades, many excellent results in performance analysis have been obtained, and successful examples abound. However, exploring special features of queueing systems directly in performance optimization still seems to be a territory not very well cultivated. Recent progresses of perturbation analysis (PA) and sensitivity-based optimization provide a new perspective of performance optimization of queueing systems. PA utilizes the structural information of queueing systems to efficiently extract the performance sensitivity information from a sample path of system. This paper gives a brief review of PA and performance optimization of queueing systems, focusing on a fundamental concept called perturbation realization factors, which captures the special dynamic feature of a queueing system. With the perturbation realization factors as building blocks, the performance derivative formula and performance difference formula can be obtained. With performance derivatives, gradient-based optimization can be derived, while with performance difference, policy iteration and optimality equations can be derived. These two fundamental formulas provide a foundation for performance optimization of queueing systems from a sensitivity-based point of view. We hope this survey may provide some inspirations on this promising research topic.     

Using Cumulant Functions in Queueing Theory

A new approach for obtaining the transient solution for the first and second moments of the system size in a finite capacity M/M/1 queueing systems is developed. The approach uses the cumulant generating function which has previously been used in the analysis of compartmental models but has not been used to analyze queueing systems.     

Perturbation analysis and optimization of queueing networks

A new time-domain-based approach is developed in this paper for the perturbation analysis of queueing networks. We show that, by observing a single sample path realization of the network trajectory, we can derive sensitivity information of the throughput of the system with respect to various parameters. This information can then be used for the optimization of queueing networks. Numerous experiments as well as analytical results demonstrating the validity of this new approach are given and discussed.     

Stochastic design optimization of asynchronous flexible assembly systems

This paper presents the application of the stochastic quasigradient method (SQG) of Ermoliev and Gaivaronski to the performance optimization of asynchronous flexible assembly systems (AFAS). These systems are subject to blocking and starvation effects that make complete analytic performance modeling difficult. A hybrid algorithm is presented in this paper which uses a queueing network model to set the number of pallets in the system and then an SQG algorithm is used to set the buffer spacings to obtain optimal system throughput. Different forms of the SQG algorithm are examined and the specification of optimal buffer sizes and pallet numbers for a variety of assembly systems with ten stations are discussed. The combined Network-SQG method appears to perform well in obtaining a near optimal solution in this discrete optimization example, even though the SQG method was primarily designed for application to differentiable performance functionals. While a number of both theoretical and practical problems remain to be resolved, a heuristic version of the SQG method appears to be a reasonable technique for analyzing optimization problems for certain complex manufacturing systems.     

Capacity planning in manufacturing and computer networks

This paper addresses capacity planning in manufacturing and computer networks. More specifically, given a manufacturing or computer system modeled as a network of queues, we consider the minimum cost selection of capacity levels from a discrete set of choices such that a single system performance constraint is satisfied. We focus on settings where the cost of obtaining capacity is a concave function, allowing fixed charges and economies of scale to be handled. To solve this class of capacity planning problems, we present a branch and bound algorithm that globally minimizes a concave cost function over a single convex nonlinear performance constraint and lower and upper bounds on the discrete capacity variables. We also present reoptimization procedures that allow the subproblems to be solved more efficiently. Computational results with the algorithm are reported.     

Exact asymptotic formulae of the stationary distribution of a discrete-time two-dimensional QBD process

We propose an analytically tractable approach for studying the transient behavior of multi-server queueing systems and feed-forward networks. We model the queueing primitives via polyhedral uncertainty sets inspired by the limit laws of probability. These uncertainty sets are characterized by variability parameters that control the degree of conservatism of the model. Assuming the inter-arrival and service times belong to such uncertainty sets, we obtain closed-form expressions for the worst case transient system time in multi-server queues and feed-forward networks with deterministic routing. These analytic formulas offer rich qualitative insights on the dependence of the system times as a function of the variability parameters and the fundamental quantities in the queueing system. To approximate the average behavior, we treat the variability parameters as random variables and infer their density by using ideas from queues in heavy traffic under reflected Brownian motion. We then average the worst case values obtained with respect to the variability parameters. Our averaging approach yields approximations that match the diffusion approximations for a single queue with light-tailed primitives and allows us to extend the framework to heavy-tailed feed-forward networks. Our methodology achieves significant computational tractability and provides accurate approximations for the expected system time relative to simulated values.     

Monotone control of queueing networks

This paper uses submodularity to obtain monotonicity results for a class of Markovian queueing network service rate control problems. Nonlinear costs of queueing and service are allowed. In contrast to Weber and Stidham [14], our monotonicity theorem considers arbitrary directions in the state space (not just control directions), arrival routing problems, and certain uncontrolled service rates. We also show that, without service costs, transition-monotone controls can be described by simple control regions and switching functions. The theory is applied to queueing networks that arise in a manufacturing system that produces to a forecast of customer demand, and also to assembly and disassembly networks.     

Multi-objective performance improvements of general finite single-server queueing networks

Optimizing the performance of general finite single-server acyclic queueing networks is a challenging problem and has been the subject of many studies. The version of the optimization problem treated here considers the minimization of the buffer areas and the service rates simultaneously with the maximization of the throughput. These are conflicting objectives, and the most appropriate methodology appears to be a multi-objective methodology. In fact, algorithms have previously been proposed, and the aim here is to show that the use of a mixed methodology can occasionally improve solutions without a significant increase in the computational costs. This paper shows that improvements in throughput can be achieved through a solution of a type of stochastic knapsack problem, which consists of redistributing the buffer spaces between the lines while preserving the overall capacity using a simulated annealing algorithm; that is, one objective is improved (the throughput) without worsening the other (the overall allocated capacity). A set of computational experiments are presented to demonstrate the effectiveness of the proposed approach. Additionally, some of the insights presented here may help scientists and practitioners in finite single-server queueing network planning.     

Queueing network modelling of flexible manufacturing system using mean value analysis

In the present investigation, we develop queueing model for the performance prediction of flexible manufacturing systems (FMSs) with a multiple discrete material-handling devices (MHD). An iterative method has been suggested using mean value analysis (MVA) for the state-dependent routing. Two queueing network models are considered to determine the material-handling device interference. In the first one, we model the interference from the MHD by inflating the station service times but neglect queueing at the MHD. In another network, the queueing for the MHD is taken into consideration. The performance of FMS configuration is obtained by iterating between two networks. The suggested algorithms demonstrate better results than the algorithm used by earlier workers for single MHD. Some performance indices viz. throughput, mean service time, mean waiting time, etc. are obtained. Numerical results are provided to highlight the effect of the system parameters on performance indices, which are further evaluated by using neuro-fuzzy controller system to validate the tactability of soft computing approach.     

Random-Direction Optimization Algorithms with Applications to Threshold Controls

This work develops a class of stochastic optimization algorithms. It aims to provide numerical procedures for solving threshold-type optimal control problems. The main motivation stems from applications involving optimal or suboptimal hedging policies, for example, production planning of manufacturing systems including random demand and stochastic machine capacity. The proposed algorithm is a constrained stochastic approximation procedure that uses random-direction finite-difference gradient estimates. Under fairly general conditions, the convergence of the algorithm is established and the rate of convergence is also derived. A numerical example is reported to demonstrate the performance of the algorithm.     

Bounding blocking probabilities and throughput in queueing networks with buffer capacity constraints

We propose a new technique for upper and lower bounding of the throughput and blocking probabilities in queueing networks with buffer capacity constraints, i.e., some buffers in the network have finite capacity. By studying the evolution of multinomials of the state of the system in its assumed steady state, we obtain constraints on the possible behavior of the system. Using these constraints, we obtain linear programs whose values upper and lower bound the performance measures of interest, namely throughputs or blocking probabilities. The main advantages of this new technique are that the computational complexity does not increase with the size of the finite buffers and that the technique is applicable to systems in which some buffers have infinite capacity. The technique is demonstrated on examples taken from both manufacturing systems and communication networks. As a special case, for the M/M/s/s queue, we establish the asymptotic exactness of the bounds, i.e., that the bounds on the blocking probability asymptotically approach the exact value as the degree of the multinomials considered is increased to infinity. This revised version was published online in June 2006 with corrections to the Cover Date.     

Tandem queueing networks with neighbor blocking and back-offs

We introduce a novel class of tandem queueing networks which arise in modeling the congestion behavior of wireless multi-hop networks with distributed medium access control. These models provide valuable insight in how the network performance in terms of throughput depends on the back-off mechanism that governs the competition among neighboring nodes for access to the medium. The models fall at the interface between classical queueing networks and interacting particle systems, and give rise to high-dimensional stochastic processes that challenge existing methodologies. We present various open problems and conjectures, which are supported by partial results for special cases and limit regimes as well as simulation experiments.