Han, Chingping (Jim)

The future of manufacturing industry is extending from the high profile systems like CAD and CAM, to a system that provides real time knowledge towards the manufacturing process. Computer Aided Process Planning (CAPP) is such a system. The computer provides the CAPP system a means to store and search previously developed process plans to find a similar process for a current part. In this research a CAPP application was developed using dBase IV, to provide a link between manufacturing process plans and engineering data. The relational database used provides an effective way to manage and manipulate manufacturing process information. The application uses engineering and manufacturing data from Curt G. Joa, Inc., to verify its operation. The development of the application, its corresponding computer programs, and literature survey are discussed in this thesis.

In this study we developed a supply chain contract model for multiple scheduling period with dynamic demand patterns of stochastic nature, and with elastic price structures. The model presented here combined and enhanced several supply chain contract models developed previously. It is unique in that it considered multiple periods, dynamic, stochastic and price-elastic demand patterns, and flexible order quantities. Using a linear demand price-elastic relation and normal distribution pattern, optimal solutions for minimum cost, maximum profit, price structure, and order policies for the entire supply chain were derived. Sensitivity analyses performed in this study gave a better understanding of relative importance of various system variables.

One compelling reason for converting to reusable containers is the conservation of natural resources. Other than the attractiveness of a potential cost savings, it could provide the firm with a good public relations image worth even more. Many firms are working toward being ISO (international Standards Organization) 14001 certified. ISO 14001 includes in its guidelines the need to reduce pollution. With this in mind, there is a need for better tools to predict cost when converting from disposable to reusable containers. This study provides a tool to estimate conversion cost based on a mathematical model that captures the important elements of cost. It is also based on inventory control methods, design of experiments and a simulation model with animation that captures these cost elements as well. Some guidelines that will help sell a reusable container project to upper management, together with a discussion on implementation logistics, is presented as well.

The Material Requirement Planning System (MRP) is a key module in the Enterprise Resources Planning (ERP) systems. Finding better ways of developing systems leads to a robust MRP system and provides prototypes which can be expanded to other modules of the ERP systems. This thesis will provide an object-oriented model of a MRP system which was created and presented by using the Object Modeling Technique (OMT). The model consisted of three parts, the object model, dynamic model, and functional model. An innovative way to handle the static bill of materials (BOM) data with a more flexible dynamic item class was developed. To demonstrate the flexibility, extendability, and reusability of the object-oriented MRP model, two major business change scenarios were applied to the model. Minor changes to the design were required to accommodate major changes on the functionality of the MRP system.

This thesis presents a model designed to optimize the allocation of corporate resources required for the success of a product in the marketplace. The product development resources used in the model are: market research, applied research, product design, cost reduction and advertising. The key goals of this thesis are to provide industry with a usable tool: (1) Implement strategic plans through effective budgeting; (2) Optimize both short and long term profits; (3) Evaluate the impact of resource inter-dependencies; (4) Enable accountability that leads to goal achievement and checks unnecessary growth; (5) Remove much of the negative political and emotional variability; (6) Easily adapt to internal and external changes; (7) Output a specific allocation for each resource as a percentage of sales; (8) Output an estimate of future profitability. Genetic Algorithms are particularly well suited for this application because an exact optima is not required and the search space can be extremely large, complex, and non-linear.

Manufacturing Resource Planning systems are functionally complex systems. In providing effective resource management tools, they support the integration of a wide variety of complex functions. These systems also undergo frequent changes as business needs change. For these reasons, analysis techniques which provide methods to create clear, flexible systems must be sought. Object oriented analysis is such a technique. This thesis presents the development of an object oriented model for a Manufacturing Resource Planning system (MRPII). It will be shown that the use of objects and object oriented techniques to model complex systems such as MRPII results in system models which are more easily understood and more flexible to change than other more conventional representations. Future research may include the formal design and implementation of the model. The flexibility of the implemented system could then be compared to the level of flexibility of a non-object based system.

Vision systems have been widely used for parts inspection in electronics assembly lines. In order to improve the overall performance of a visual inspection system, it is important to employ an efficient object recognition algorithm. In this thesis work, a genetic algorithm based correlation algorithm is designed for the task of visual electronic parts inspection. The proposed procedure is composed of two stages. In the first stage, a genetic algorithm is devised to find a sufficient number of candidate image windows. For each candidate window, the correlation is performed between the sampled template and the image pattern inside the window. In the second stage, local searches are conducted in the neighborhood of these candidate windows. Among all the searched locations, the one that has a highest correlation value with the given template is selected as the best matched location. To apply the genetic algorithm technique, a number of important issues, such as selection of a fitness function, design of a coding scheme, and tuning of genetic parameters are addressed in the thesis. Experimental studies have confirmed that the proposed GA-based correlation method is much more effective in terms of accuracy and speed in locating the desired object, compared with the existing Monte-Carlo random search method.

Hospital Emergency Departments are forced to operate under many unknown conditions. Therefore, determining the appropriate staffing requirements of an emergency department can be a difficult task. The objective herein is to develop a procedure for assembling a simulation model, to be used as a tool for determining adequate resource requirements for various situations encountered in an emergency department of a hospital. The model will consider variables such as patient arrival rates, types of injuries, duration of treatment for each specific injury, available resources, etc. It is expected that the use of such a model will help emergency department managers plan for adequate resources to meet the needs of the community being served.

The crucial goal of enhancing industrial productivity has led researchers to look for robust and efficient solutions to problems in production systems. Evolving technologies has also, led to an immediate demand for algorithms which can exploit these developments. During the last three decades there has been a growing interest in algorithms which rely on analogies to natural processes. The best known algorithms in this class include evolutionary programming, genetic algorithms, evolution strategies and neural networks. The emergence of massively parallel systems has made these inherently parallel algorithms of high practical interest. The advantages offered by these algorithms over other classical techniques has resulted in their wide acceptance. These algorithms have been applied for solving a large class of interesting problems, for which no efficient or reasonably fast algorithm exists. This thesis extends their usage to the domain of production research. Problems of high practical interest in the domain of production research are solved using a subclass of these algorithms i.e. those based on the principle of evolution. The problems include: the flowpath design of AGV systems and vehicle routing in a transportation system. Furthermore, a Genetic Based Machine Learning (GBML) system has been developed for optimal scheduling and control of a job shop.

This thesis presents a mathematical model for the sensitivity analysis of machine cell formation. Computer programs in C were developed. A statistical simulation model is developed to test and verify the mathematical model. The data for machine cell formation in cellular manufacturing is organized in a machine component chart representing the machining requirements of parts in the product mix. The existing machine cell formation models treat the product mix as deterministic. To study the probabilistic nature of the cellular manufacturing, a sensitivity analysis model is presented. The model optimizes the formation of intercellular material handling cost for the machine cell within the constrains of the probability of the product mixture. The results of the mathematical model is compared with the results of the simulation model. It shows that the probabilistic product mix has a influence on the efficiency of the machine cell and the associated total cost.