This dissertation deals with the problem of locating inspection stations in two different scenarios: (1) in a manufacturing environment, and (2) in a network involving the transportation of hazardous materials. In the manufacturing environment problem, the location of specialized inspection stations in a serial system and in a serial-parallel system is considered. These location problems have the special characteristic that the demand for the facilities being located is from prespecified link flows and their properties. When defects are introduced in a serial manufacturing system, objectives such as minimization of total cost, maximization of yield, and minimization of undetected faulty units are all viable optimization criteria. The dissertation develops several new models for locating specialized inspection stations using such objectives. Cases considered include (1) when inspection/rework stations are to be located, (2) when inspection stations are already located but their operating modes (rework or scrap) are to be determined, and (3) when both locations and operating modes are to be determined. Exact and/or heuristic methods of solutions for these models are developed and some computational experience is reported. Serial lines have the disadvantage that the product flow is easily disrupted by a single machine malfunction anywhere on the line. One way of improving system performance is to provide some degree of redundancy at the machine level using serial-parallel systems. Extensions of some of the serial-system results to the serial-parallel systems are also presented. In the hazardous material transportation setting, the regulating agencies problem of making decisions on where to inspect trucks on the underlying transportation network is considered. In the case when each inspection station's capacity is limited, the problem can be formulated as a new capacitated facility location problem. Some heuristics and an exact branch-and-bound procedure to solve this problem are developed and some computational experience is reported.

Many industries are in the process of adopting or considering adoption of the Just-In-Time philosophy and techniques. Most of these firms are not characterized by the type of environment which is required for transferring the philosophy, such as stable manufacturing factors. The adverse effect of environmental uncertainties on the operating phase of Just-In-Time systems in general and the performance of Kanban systems in particular has been well documented. This dissertation proposes a framework for designing robust Kanban systems, by adopting the product and process design approach pioneered by Genichi Taguchi. A robust Kanban systems is one whose performance is insensitive to environmental uncertainties. As part of our investigation, we illustrate how frequency domain experiments can be used in order to investigate the effect of noise factors and their interactions on the performance of a Kanban system. The technique allows us to examine simultaneously many factors that previous researchers have studied separately, and so explore and identify important interaction effects. We develop a general framework for designing a robust Kanban system that can be applied to different production configurations, and illustrate the robust design process via a simulation study of a three-stage manufacturing line.

The most promising switching technique for B-ISDN (broadband integrated service digital network) is the ATM (asynchronous transfer mode). In an ATM network, all information, data, voice and video, is packetized and divided into fixed length data blocks called cells. The cells from different connections are carried through a multiplexer, and asynchronously transmitted through the network. Statistical multiplexing of cells allows the possible reduction of the bandwidth assigned to each single source. That increased flexibility with respect to the bandwidth requirement provides a chance for better, more economical utilization of the network resources. On the other hand, severe network congestion can occur when a large number of traffic sources become active simultaneously. Since most traffic sources in ATM networks are bursty, some congestion control must be applied to each source in order to maintain the required GOS (grade of service) and provide fairness among the users. We introduce the discrete batch Markovian arrival process, which is a versatile and tractable class of Markov renewal processes. This class of processes provides a very powerful modeling tool. The Palm measure, variance time curve, asymptotic normality of the counts are derived. The interarrival time distribution for the single arrivals case are discussed. We also address some issues related to the simulation of this class of processes. Two traffic shaping, or smoothing schemes are investigated in this dissertation: jumping windows with regular placement and an input rate control model, introduced by Ohta et al. (21). The discrete Markovian arrival process with single arrivals serves as the model for the arrival process. In the first model, analytical expressions for the loss probability, packet delay and the interarrival times for the shaped process in steady state are derived. The second model leads to a highly degenerate partitioned Markov chain of QBD (Quasi-Birth-and-Death) type. Special algorithms involving matrices of lower order are obtained by exploiting the special structure of the Markov chain. Some performance measurements are derived. The algorithmic implementation of these results is also discussed. Finally, we examine some specific examples, applying both the analytical results and simulation, to demonstrate the effectiveness of the two traffic shaping schemes.

The ultimate goal of this work is to provide a general global optimization method. Due to the difficulty of the problem, the complete task is divided into several sections. These sections can be collected into a modeling phase followed by a global minimization phase. Each of the various sections draws from different engineering fields. What this work suggests is an interface and common ground between these fields. The modeling phase of the procedure consists of converting a general problem into a given formulation using a particular type of neural network. The architecture required for the neural network forms a new class: the pseudo multilayer neural network. It is introduced and compared to more classical neural network architectures such as the regular multilayer neural network. However, a key difference between these two classes is that the behavior of the usual multilayer network has to be programmed via iterative training, while an extremely efficient direct procedure is given here to synthesize the pseudo multilayer neural network. Therefore any initial problem can be systematically converted into a pseudo multilayer network without going through the undesired programming steps such as the backpropagation rule. The second phase of the work consists of translating the initial global optimization problem into the global minimization of a target function related to the neural network model. Systematic procedures are again given here. The last phase consists of globally minimizing the target function. This is done via the so-called DC programming technique where DC stands for "Difference of Convex". The pseudo multilayer was created such that it can systematically be converted into a DC formulation, and therefore be compatible with DC programming. A translation procedure to go from the pseudo multilayer neural network model to the DC formulation is given. When a DC program is applied to this last formulation, the resulting solution can be directly mapped to the global minimum of the target function previously defined, thereby producing the global optimal solution of the neural network modeling the initial problem. Therefore, the optimal solution of the original problem is known as well.

Stochastic linear programs are linear programs in which some of the problem data are random variables. The particular kind of programs that we study belong to the recourse model. Under this model, some decisions are postponed until better information becomes available (e.g., an outcome of a random variable is realized), while other decisions must be made 'here and now.' For example, in a telecommunication network planning problem, decisions regarding the addition of network capacity have to be made before knowing customer demand (i.e., 'here and now'). Once the demand is realized, efficient usage of the network can then be determined. This work explores algorithms for the solution of such programs: stochastic linear programs with recourse. The algorithms investigated can be described as decomposition based cutting plane methods in which the cuts are estimated from random samples. Moreover, the algorithms all use the incremental sampling plan inherent to the Stochastic Decomposition (SD) algorithm developed by Higle and Sen in 1991. Our study includes both two stage and multistage programs. For the solution of two stage programs, we present the Conditional Stochastic Decomposition (CSD) algorithm, a multicut version of the SD algorithm. CSD is most suitable for situations in which data are difficult to obtain and may be computationally intense. Because of this potential intensity, we explore algorithms which require less computational effort than CSD. These algorithms combine features of both CSD and SD and are referred to as hybrid algorithms. Following our exploration of these algorithms for two stage problems, we next explore an extension of the SD algorithm that can be used for multistage problems with stagewise independent random variables. For the sake of notational brevity, our technical development is centered around the three stage case, although the extension to multistage problems is straightforward. Under mild conditions, convergence results similar to those found in the two stage algorithms hold. Multistage stochastic decomposition is currently a largely uncharted area. Our research represents the first major effort in this direction.

No description available.

<p> Through this dissertation, the researcher demonstrated validity of a new individual level construct titled <i>perceived business process knowledge</i> (PBPK) and an associated measurement instrument. The concepts and definition of knowledge and business processes from past works were integrated to define PBPK. Factors found in empirically tested information systems models were used to identify five PBPK sub-dimensions: 1) Perceived Business Process Input Knowledge, 2) Perceived Business Process Resources Knowledge, 3) Perceived Business Process Activities Knowledge, 4) Perceived Business Process Constraints Knowledge, and 5) Perceived Business Process Output Knowledge. A sixth sub-dimension, Perceived Business Process Monitoring Knowledge, surfaced in the field research phase.</p><p> The sub-dimensions and focal construct were tested for criterion and predictive validity using SmartPLS and demonstrated high reliability and validity under two samples. All indicators have loadings greater than 0.860 (p &lt; 0.000). PBPK items demonstrated Cronbach Alpha greater than 0.85 and Composite Reliability > 0.95 on both data collections. Focal construct and sub-dimensions demonstrated convergent and discriminant validity; Average Variance Extracted (AVE) greater than 0.5, and square root of the AVE greater than associated correlation with all other constructs.</p><p> The first data collection consisted of representatives from 21 industries; the criterion and predictive validity relationships produced almost acceptable but low R² (> 0.21), medium f², (> 0.27) and moderate path coefficients (> 0.46, p &lt; 0.000) on the smallest predictor to dependent variable relationship. The second data collection consisted of representatives from 3 industries; the criterion and predictive validity relationships were a slightly better; the results demonstrated acceptable but low R² (> 0.29), large f² (> 0.41), and moderate path coefficients (> 0.53, p &lt; 0.000) on the smallest predictor to dependent variable relationship. </p><p> The study also successfully demonstrated that different groups of individuals report different levels of PBPK. These tests were conducted using one-way ANOVA.</p><p> The constructs and associated measurement instrument are expected to increase the predictive ability of Technology Acceptance Models.</p>

The Defense Language Institute (DLI) offers 23 beginning language courses and in 2004 began to provide a smaller class size for these courses. Restrictions on when classes can begin and a limited number of instructors prevent all students from being trained in a smaller class. This thesis develops integer linear programs (ILPs) that generate schedules for all student classes and maximize the number of smaller class starts for a given number of instructors. Secondary scheduling goals include avoiding weekly changes to instructor levels and scheduling preferences such as the number of classes to start simultaneously. The ILPs solve in less than one minute and offer a significant improvement in the number of students that may be trained in the smaller class size. Computational results using real data for the Arabic, Chinese-Mandarin, and Persian-Farsi courses verify the ILPs find feasible multiyear schedules that incorporate the DLI's scheduling preferences while exceeding the DLI's published schedule results. For example, the ILPs find schedules for Arabic that train 8%, 34% and 76% of students in the smaller class in 2006, 2007, and 2008, whereas DLI's manual schedules at best can train 8%, 7% and 64%.

Discrete Event Simulation (DES) is one of the most widely used methodologies for Operations Research (OR) modeling and analysis. However, designing and implementing DES can be a time-consuming and error-prone task. This thesis designed, implemented and evaluated a tool, the Event Graph Graphical Design Tool (EGGDT), to help OR analystsin the design, implementation, and maintenance of DES reducing the development and debugging times. The Unified Modeling Language was used to document the development of the EGGDT, which was programmed in Java using J2D and Swing. Human Factors techniques were employed to help in the design process and to evaluate the final prototype of the EGGDT. During the design process, two formative experiments were performed to evaluate the Graphical User Interface design decisions. A final summative experiment was done to test if the potential users consider the tool a useful means to develop OR simulations. Participants of the experiments agreed that tools like the EGGDT are an essential instrument when developing simulation. / Spanish Navy author

"The Defense Language Institute Foreign Language Center (DLIFLC) trains students in various foreign languages and dialects for the Department of Defense (DOD). The majority of students are firstterm enlistees in the basic program. This study uses classification trees and logistic regression to understand the military, academic and personal characteristics that influence first-term success after successfully completing DLIFLC training. Success was defined as completing a firstterm enlistment contract and maintenance of language proficiency. DLIFLC management was interested in the difference in success for individuals that graduated DLIFLC via the different training pipelines. Students graduate by completing the program as originally assigned, or by recycling, relanguaging or taking DLPT enhancement training multiple times and in multiple combinations due to various academic, administrative or other reasons. 63% of students graduated. Only 45% of those that graduated were successful post-DLIFLC. Results identified several factors influential in predicting success; the factors were service affiliation, contract lengths and gender. Training pipelines were slightly influential. Individuals in the Army had the worst odds of success. Contract lengths greater than four years had lower odds of success. Males had higher odds of success than females." p. i.