David D. Yao

A Supply Network Model with Base-Stock Control and Service Requirements

We develop a supply network model that takes as input the bill of materials, the (nominal) lead times, the demand and cost data, and the required customer service levels. In return, the model generates the base-stock level at each store--the stocking location for a part or an end-product, so as to minimize the overall inventory capital throughout the network and to guarantee the customer service requirements. The key ingredient of the model is a detailed, albeit approximate, analysis of theactual lead times at each store and the associated demand over such lead times, along with a characterization of the operation at each store via an inventory-queue model. The gradients are derived in explicit forms, and a conjugate gradient routine is used to search for the optimal solution. Several numerical examples are presented to validate the model and to illustrate its various features.

A queueing network model for semiconductor manufacturing

We develop an open queueing network model for rapid performance analysis of semiconductor manufacturing facilities. While the use of queueing models for performance evaluation of manufacturing systems is not new, our approach differs from others in the detailed ways in which we model the different tool groups found in semiconductor wafer fabrication, as well as the way in which we characterize the effect of rework and scrap on wafer lot sizes. As an application of the model, we describe a method for performing tool planning for semiconductor lines. The method is based on a marginal allocation procedure which uses performance estimates from the queueing network model to determine the number of tools needed to achieve a target cycle time, with the objective being to minimize overall equipment cost.

Queue Spillovers in Transportation Networks with a Route Choice

This paper examines different traffic phenomena that occur when drivers have to navigate a network in which queues backup past diverge intersections. In particular, it looks at the bottleneck capacity of a network and how reducing it to below a critical level can resume the attractiveness of an alternative route, thus preventing the system from reaching the saturation point. The author concludes that the time-dependent traffic assignment problem with physical queues is chaotic in nature and that it may be impossible to obtain input data with the required accuracy to make reliable predictions of cumulative output flows on severely congested networks.

Inventory-Service Optimization in Configure-to-Order Systems

This study is motivated by a process-reengineering problem in personal computer (PC) manufacturing, i.e., to move from a build-to-stock operation that is centered around end-product inventory towards a configure-to-order (CTO) operation that eliminates endproduct inventory. In fact, CTO has made irrelevant the notion of preconfigured machine types and focuses instead on maintaining the right amount of inventory at the components. CTO appears to be the ideal operational model that provides both mass customization and a quick response time to order fulfillment. To quantify the inventory-service trade-off in the CTO environment, we develop a nonlinear optimization model with multiple constraints, reflecting the service levels offered to different market segments. To solve the optimization problem, we develop an exact algorithm for the important case of demand in each market segment having (at least) one unique component, and a greedy heuristic for the general (nonunique component) case. Furthermore, we show how to use sensitivity analysis, along with simulation, to fine-tune the solutions. The performance of the model and the solution approach is examined by extensive numerical studies on realistic problem data. We present the major findings in applying our model to study the inventory-service impacts in the reengineering of a PC manufacturing process.

Performance Analysis and Optimization of Assemble-to-Order Systems with Random Lead Times

We study a single-product assembly system in which the final product is assembled to order whereas the components (subassemblies) are built to stock. Customer demand follows a Poisson process, and replenishment lead times for each component are independent and identically distributed random variables. For any given base-stock policy, the exact performance analysis reduces to the evaluation of a set ofM/ G/8 queues with a common arrival stream. We show that unlike the standardM/ G/8 queueing system, lead time (service time) variability degrades performance in this assembly system. We also show that it is desirable to keep higher base-stock levels for components with longer mean lead times (and lower unit costs). We derive easy-to-compute performance bounds and use them as surrogates for the performance measures in several optimization problems that seek the best trade-off between inventory and customer service. Greedy-type algorithms are developed to solve the surrogate problems. Numerical examples indicate that these algorithms provide efficient solutions and valuable insights to the optimal inventory/service trade-off in the original problems.

Optimal peer selection for P2P downloading and streaming

In a P2P system, a client peer may select one or more server peers to download a specific file. In a P2P resource economy, the server peers charge the client for the downloading. A server peers price would naturally depend on the specific object being downloaded, the duration of the download, and the rate at which the download is to occur. The optimal peer selection problem is to select, from the set of peers that have the desired object, the subset of peers and download rates that minimizes cost. In this paper we examine a number of natural peer selection problems for both P2P downloading and P2P streaming. For downloading, we obtain the optimal solution for minimizing the download delay subject to a budget constraint, as well as the corresponding Nash equilibrium. For the streaming problem, we obtain a solution that minimizes cost subject to continuous playback while allowing for one or more server peers to fail during the streaming process. The methodologies developed in this paper are applicable to a variety of P2P resource economy problems.

Order Fill Rate, Leadtime Variability, and Advance Demand Information in an Assemble-to-Order System

We study an assemble-to-order system with stochastic leadtimes for component replenishment. There are multiple product types, of which orders arrive at the system following batch Poisson processes. Base-stock policies are used to control component inventories. We analyze the system as a set of queues driven by a common, multiclass batch Poisson input, and derive the joint queue-length distribution. The result leads to simple, closed-form expressions of the first two moments, in particular the covariances, which capture the dependence structure of the system. Based on the joint distribution and the moments, we derive easy-to-compute approximations and bounds for the order fulfillment performance measures. We also examine the impact of demand and leadtime variability, and investigate the value of advance demand information.

Monotone structure in discrete-event systems

Some Basic Concepts. Antimatroid Structure: Monotonicity. Lattice Structure: Convexity and Concavity. Links to Other Models. Monotone Optimal Control. Subadditivity and Stability. Association and Optimal Coupling. Perturbation Analysis. Index.

Multiclass queueing systems: polymatroidal structure and optimal scheduling control

In many multiclass queueing systems, certain performance measures of interest satisfy strong conservation laws. That is, the total performance over all job types is invariant under any nonidling service control rule, and the total performance over any subset (say A) of job types is minimized or maximized by offering absolute priority to the types in A over all other types. We develop a formal definition of strong conservation laws, and show that as a necessary consequence of these strong conservation laws, the state space of the performance vector is a (base of a) polymatroid. From known results in polymatroidal theory, the vertices of this polyhedron are easily identified, and these vertices correspond to absolute priority rules. A wide variety of multiclass queueing systems are shown to have this polymatroidal structure, which greatly facilitates the study of the optimal scheduling control of such systems. When the defining set function of the performance space belongs to the class of generalized symmet...

Extended-Enterprise Supply-Chain Management at IBM Personal Systems Group and Other Divisions

In 1994, IBM began to reengineer its global supply chain. It wanted to achieve quick responsiveness to customers with minimal inventory. To support this effort, we developed an extended-enterprise supply-chain analysis tool, the Asset Management Tool (AMT). AMT integrates graphical process modeling, analytical performance optimization, simulation, activity-based costing, and enterprise database connectivity into a system that allows quantitative analysis of extended supply chains. IBM has used AMT to study such issues as inventory budgets, turnover objectives, customer-service targets, and new-product introductions. We have implemented it at a number of IBM business units and their channel partners. AMT benefits include over