Browsing by Author "Dr. Russell E. King, Committee Member"

Now showing 1 - 8 of 8

Autonomous Solution Methods for Large-Scale Markov Chains
(2002-08-19) Barge, Walter S., II; Dr. Robert St. Amant, Committee Member; Dr. Russell E. King, Committee Member; Dr. Thom J. Hodgson, Committee Member; Dr. William J. Stewart, Committee Chair
One of the roadblocks to greater application of Markov chains is that non-numerically sophisticated users possess the detailed domain knowledge needed to construct a large Markov chain but may have a difficult time deciding which numerical solution method might be best suited to their applications. A realistic Markov chain model can easily contain hundreds of thousands of states, yet users may severely restrict their models to keep them small enough to fit within the constraints of certain software packages or solution methods. Even after selecting a solution method, implementation details imposed by compact storage schemes and the nature of the solution method itself may pose additional barriers. By making judgments about the Markov chain, an experienced researcher or practitioner can sometimes propose a solution technique in a short amount of time. This research examines methods to obtain a proposed solution technique without the services of an expert and with little or no intervention from the novice user. We take advantage of information readily available in the Markov chain to aid in the selection and execution of a solution method. We demonstrate a computer tool with a graphical user interface (GUI) and embedded expert system to make large-scale Markov chain analysis more accessible. The computer tool receives a user's Markov chain, examines the chain, determines its primary characteristics, and then gives the user useful information and recommendations about how to analyze the model. This can be done without the user being an expert in the various solution techniques and their respective areas of applicability.
Design of a Public Logistics Network.
(2004-05-20) Bansal, Amogh; Dr. Michael G. Kay, Committee Chair; Dr. Russell E. King, Committee Member; Dr. Negash Medhin, Committee Member
This thesis presents a design for a public logistics network (PLN) covering the continental United States. A systematic approach is developed to determine the initial national road network using only the Interstate highways and part of the U.S. highways. Various heuristics are developed for generating the underlying road network. In two of the heuristics, every Interstate node is used and then U.S. highway nodes are added to the road network to supplement the Interstate nodes. Another heuristic generates the network by directly joining the roads (based on shortest time routes) between the cities of more than certain population. The results from testing show that the later heuristic performs better than the former ones. This road network is then used for developing the PLN by selecting some of its nodes as the locations for distribution centers (DCs). The PLN is then developed by removing and adding arcs in a reduced network obtained by Delaunay triangulation of the selected DC nodes in the underlying road network. Given the number and location of DCs on this underlying network, the minimum average transport time for a package is used as the criterion to compare alternative PLN designs. The package demand used to determine the minimum average transport time is proportional to the population at each five-digit zip code centroid surrounding each DC. Effects of different parameters on the design of the PLN are studied. Finally, a genetic algorithm is used to get the optimal public logistics network for the entire U.S.
Game Theoretic Approach to Supply Chain Management
(2003-07-08) Dai, Yue; Dr. Russell E. King, Committee Member; Dr. Henry L.W. Nuttle, Committee Co-Chair; Dr. Shu-Cherng Fang, Committee Chair; Dr. Xiuli Chao, Committee Co-Chair
This dissertation studies the competitive behavior of firms in supply chain management and revenue management contexts. A game theoretic approach is employed. We analyze capacity allocation and pricing strategies and derive equilibrium solutions for multiple competing firms. We also study channel coordination mechanisms to bring the competing firms together for chain-wide optimality and conduct sensitivity analysis of equilibrium solutions. First we consider a single-period distribution system with one supplier and two retailers. When a stockout occurs at one retailer the customer may go to the other retailer. The supplier may have infinite or finite capacity. In the latter case, if the total quantity ordered (claimed) by the retailers exceeds the supplier's capacity, an allocation policy is invoked to assign the capacity to the retailers. We show that a unique Nash equilibrium exists when the supplier has infinite capacity. While, when the capacity is finite, a Nash equilibrium exists only under certain conditions. For the finite capacity case, we also use the concept of Stackelberg game to develop optimal strategies for both the leader and the follower. In addition to the decentralized inventory control problem, we study the centralized inventory control problem and obtain the optimal allocation that maximizes the expected profit of the entire supply chain. We also design perfect coordination mechanisms, i.e., a decentralized cost structure resulting in a Nash equilibrium with chain-wide profits equal to those achieved under a fully centralized system. As an extension to the capacity allocation models above, we then consider two firms where each firm has a local store and an online store. Customers may shift among these stores upon encountering a stockout. Each firm makes the capacity allocation decision to maximize its profit. We consider two scenarios of a single-product single-period model and derive corresponding existence and stability conditions of a Nash equilibrium. We then conduct sensitivity analysis of the equilibrium solution with respect to price and cost parameters. Finally we extend the results to a multi-period model in which each firm decides its total capacity and allocates this capacity between its local and online stores. A myopic solution is derived and shown to be a Nash equilibrium solution of a corresponding sequential game. Finally, we consider the pricing strategies of multiple firms providing same service and competing for a common pool of customers in a revenue management context. The demand at each firm depends on the selling prices charged by all firms, each of which satisfies demand up to a given capacity limit. We use game theory to analyze the systems under both deterministic and general stochastic demand. We derive the existence and uniqueness conditions for a Nash equilibrium and calculate the explicit Nash equilibrium point when the demand at each firm is a linear function of price.
Linear Program Construction Using Metamodeling
(2003-04-02) Parks, Judith-Marie Tyler; Dr. Russell E. King, Committee Member; Dr. Marc-David Cohen, Committee Member; Dr. Thom J. Hodgson, Committee Chair; Dr. Robert E. Young, Committee Member; Dr. Michael G. Kay, Committee Member
One of the most significant trends in data warehousing today is the integration of Metadata into data warehousing tools. A data warehouse is an area which exists on computer systems that is used for holding all of the data that an organization might possess. Metadata is 'data about data,' a dictionary and summary of data, that is held in a system catalog that is contained in a data warehouse. The purpose of this dissertation is four-fold: to show that by examining a database's system catalog, information can be extracted from it that can be used to develop a structure for building operations research applications. To show that a database's system catalog can be modified to hold the structure and the definition of a linear programming model. To show that a data table containing the linear programming model constraints can be automatically constructed based on the contents of the modified system catalog. And finally, to show that the modified system catalog can be used to guide a user in developing objective functions based on a given set of model constraints. Thus, the main contribution of the work is that it furthers the hybrid area of information technology/mathematical programming by exploiting metadata, as opposed to raw data, that is held in a data warehouse.
Process Improvement of a MRAP (Mine Resistant Ambush Protected) Vehicle Production Line
(2008-05-28) Fink, J. Kingsley, Jr.; COL Tim Trainor, Ph.D., Committee Member; Dr. Steven D. Jackson, Committee Member; Dr. Thom J. Hodgson, Committee Chair; Dr. Russell E. King, Committee Member
Production Scheduling in Knitted Fabric Dyeing and Finishing: A Case Approach
(2005-04-07) Laoboonlur, Preecha; Dr. Henry L. W. Nuttle, Committee Member; Dr. Russell E. King, Committee Member; Dr. Thom J. Hodgson, Committee Chair; Dr. Kristin A. Thoney, Committee Member
The dyeing and finishing processes represent one of the most complicated scheduling problems existing in real production. The problem combines two difficult but challenging scheduling aspects together: a flexible job shop with sequence dependent setups. The process consists of multiple operations, which can have either single or parallel machines. Chemical and fabric pile contamination cause the sequence dependent setups. According to the business strategy of the case factory, the scheduling problem is categorized as two cases, no job priority and two-job priority classification (high and low). The scheduling objective is to minimize maximum lateness, Lmax. The fundamental structure used for solving the dyeing and finishing scheduling problem is the Virtual Factory plus family scheduling. The Virtual Factory is a simulation based job shop scheduling system developed at North Carolina State University. The scheduling heuristic used in the Virtual Factory is developed based on family scheduling. Jobs are grouped into families and then families are scheduled. The schedule is accomplished by switching the positions and splitting the family members. This dissertation intends to solve the real problem. Scheduling problems are generated using real problem characteristics. The experimentation indicates that with the advantages of fast computation time and heuristics modified easily, the best approach is to apply several versions of a heuristic to get the best possible solution.
Protocol Design for a Public Logistics Network
(2004-12-28) Jain, Ashish; Dr. Russell E. King, Committee Member; Dr. Jennifer Blackhurst, Committee Member; Dr. Michael G. Kay, Committee Chair
A public logistics network (PLN) is envisioned to be a multi-firm distribution system where several companies work in coordination for package transport such that packages maximize their values and trucks their profit. The presence of a multi-firm environment makes it difficult to have centralized control over the system and hence the coordination among various parts of the system becomes difficult as compared to a private logistics network where a single company controls the entire system. As a result, a coordination mechanism is required such that system work on its own without any central authority making plans for the operation of the PLN. An analytical formula was developed to study the performance of a PLN. In this thesis, a simulation model of a PLN was developed and is used to determine the average waiting time for packages transported through a PLN for use in the analytical formula. A set of protocols was proposed for the trucks and packages in a PLN to enable packages that value transportation the highest to be transported by trucks that can most efficiently transport the packages. In this thesis, the set of protocols was implemented in a simulation model and the performance of the PLN was compared with and without the protocols. The performance measures used were weighted waiting time and weighted transportation time for packages transported through a PLN. It was found that there was a significant decrease in waiting time associated with the use of the protocols.
Strategic Analysis of Speed and Flexibility In Sourcing Textile Products
(2007-04-28) Hartman, Lisa M.; Dr. Jeffrey A. Joines, Committee Co-Chair; Dr. Russell E. King, Committee Member; Dr. Kristin A. Thoney, Committee Chair; Chris Moses, Committee Member
It is becoming increasingly difficult for textile companies in the United States to compete with companies around the world. Each company needs to determine ways that their company can compete on a global level. One way that textile companies can compete is by taking advantage of their proximity to the home market. This research looked at three types of garments: basic, seasonal, and fashion items and the advantages and disadvantages of sourcing these garments in four regions of the world: Domestic (manufacturers in the United States), American (manufacturers in Central and South America and the Caribbean), Far East (manufacturers in Asia other than Pakistan and India), and Pakistan/India (manufacturers in Pakistan and India). This research was divided into two parts: computer simulations and surveys. Computer simulations performed on the Sourcing Simulator™ looked at case studies and the effect of forecasting error, drift, lead times, and seasonality on the type of garments based on the criteria of service levels, gross margins, and inventory levels. The simulations were used to determine the amount of merchandise that should be initially ordered and the number of weeks of supply the retailer should carry in the store in order to meet desired service levels. From the simulations performed, it was determined that as the lead time of suppliers increases, the inventory levels that are needed in order to meet desired service levels also increase. Longer lead times require more inventory between reorders and initially to meet service levels. With longer lead times more inventory is needed to account for any variations in demand since it would take a retailer longer to receive a replenishment. Two surveys were administered, one to apparel manufacturers and one to apparel retailers. The purpose of the surveys was to collect supplemental data for the simulations performed. From the simulations, numerical data was collected. The surveys data expresses opinions, thoughts, and feelings that are to be used to better understand the significance of the simulations results. The surveys looked at the relationships between apparel manufacturers and apparel retailers and any concerns or problems experienced by members in the supply chain. The surveys also served the purpose of obtaining information about how sourcing decisions are made. The information can then be applied to future research to reflect the industry's methods of making decisions. From the apparel manufacturer's survey, the criteria that are most important when choosing a manufacturer were revealed. Also what factors are taken into consideration when choosing to use a manufacturer from a specific region were determined. The advantages and disadvantages of using a manufacturer from a specific region were also identified. From the questions about each region, it was shown what costs, concerns, and lead times are associated with each region. From the apparel retailer's survey, the criteria that are considered most important when choosing a manufacturer were found. The effect that the region of the world that the manufacturer is located has on their being chosen as a manufacturer was also studied. The survey also determined how retailers make decisions about purchasing, if retailers have replenishments on items, and how replenishments are handled. The performance measures used to evaluate a selling season were identified as well as what customer service levels retailers try to maintain.