Clark A. Mount-Campbell

Process optimization via neural network metamodeling

Abstract Queuing systems are often used to model manufacturing or communication systems. Simulation models have been the most accurate models of such systems, but when optimizing the systems it is often impractical to simulate all possible configurations to identify the “best” choice. In the absence of mathematical approximations, one may turn to simulation metamodeling. Traditional metamodeling schemes would involve a simulation model of the entire system run over some subset of the possible configurations. A regression model is then typically fit to the results of these runs and used to predict results for the many configurations that were not simulated. Our approach is to develop an artificial neural network (ANN) metamodel of components of the entire system rather than of the entire system. The entire system is then modeled by interconnecting the ANN metamodels. Specifically, we propose the following approach. An ANN is trained via simulation to act as a single “generic” queuing node. The ANN accepts input data on inter-arrival and service distributions in the form of histograms and generates a departure distribution and a performance measure. A network of queues is then modeled by interconnecting copies of the generic queuing node. To illustrate the process, a simple manufacturing system is modeled. The ANN network model is used to select the buffer sizes in front of each process that provide a desired mix of product throughput rates, while minimizing average system sojourn time.

An approach to the design of a manufacturing cell under economic considerations

Abstract A method is proposed for the system design of a manufacturing cell aiming for its profit maximization over a certain period of time. The proposed method makes use of simulation, design of experiments, regression analysis, Taguchi methods, and a profit model to generate several feasible and potentially profitable designs. A decision-maker may then choose the best alternative based on the profit value, robustness, and other practical considerations. The application of the method is illustrated with a case study where a partial design of a manufacturing cell is accomplished in a manufacturing company.

Modelling and simulation in reactive polymer processing

Modelling and simulation in reactive polymer processing have been active research areas for the past decades in academic institutions as well as within the industry. Both areas have played a key role in advancing and optimizing reactive polymer processing operations. The objective of this paper is to review the two major classifications of models used to simulate polymer processes: physics based models and empirical models. Additionally, a section on multiple criteria optimization using data envelopment analysis has been included for completeness. The work presented here helps define a decision-making framework for the creation of reactive polymer process models and for the effective selection of settings of the process variables based on these models.

The Effect of Inspector Accuracy on the Type I and Type II Errors of Common Sampling Techniques

The single sample and sequential sampling models are presented in a form which includes inspector error. The effect of inspector errors on the probability of type I and type II errors is discussed...

Decision criteria in capital budgeting under uncertainties: implications for future research

Abstract Capital budgeting decisions about physical, productive assets such as those of the manufacturing firm are studied via simulation experiments. Four criteria are investigated: net present worth (NPW); utility of NPW of cash-flows (UNC); dynamic utility of NPW (DUN), which explicitly incorporates dependence of preferences on equity position, and DUN without discounting. The latter of these is equivalent to a form of Thompson and Thuesens dynamic investment criteria. The decision procedure selects that combination of available productive assets that maximizes the value of the criterion employed in any given experimental run. Two main types of uncertainty are examined: uncertainty about project cash-flows (cash-flow uncertainty) and uncertainty due to incomplete information (basic uncertainty). Incomplete information assumes that at any particular decision time the firm has made no cash-flow estimates of future project opportunities. No significant difference is found in the performance of the firm using different criteria, provided the same discount rate is used. Cash-flow uncertainty is found to have a much larger effect than basic uncertainty. Quality of cash-flow information is found to be significant in the performance of the firm. An unexplained result is found that using no discounting with low quality of information results in statistically better performance relative to using an 8% discount rate. The implications to future research directions in capital budgeting are: (1) it seems inadvisable to focus much research effort on increasing the sophistication of project evaluation; (2) the firm should concentrate its effort on getting high quality cash-flow information of current projects; and (3) capital budgeting may benefit from a re-emphasis of research effort from project evaluation to other phases of the capital budgeting process.

MULTIPLE QUALITY CRITERIA OPTIMIZATION IN REACTIVE IN-MOLD COATING WITH A DATA ENVELOPMENT ANALYSIS APPROACH II: A CASE WITH MORE THAN THREE PERFORMANCE MEASURES

Sheet Molding Compound (SMC) is a widely utilized material to manufacture automotive exterior body panels. Compression molded SMC parts are often coated with the objective to provide additional surface properties, provide environmental protection, and/or enhance aesthetics. One of the most rapidly increasing coating methods is called In-Mold Coating (IMC) in which a liquid thermoset is injected onto the surface of the cured SMC part while this is still in the mold. Cycle time, dimensional consistency, and surface finish are among the most important performance measures (PMs) to consider in the production of SMC due to their impact in profit and quality, hence these measures are also important when using IMC. Frequently, the PMs exhibit conflicting behavior i.e. lowering the cycle time might imply decreasing the part surface quality and/or achieving a lower overall part dimensional consistency. For this reason, one must exercise especial care to identify the best compromises between the PMs along with the processing conditions that result in these best compromises. The task of finding the best compromises poses a multiple criteria optimization problem. This paper describes an application of IMC to SMC where the multiple criteria optimization problem is addressed with a non-parametric approach known as Data Envelopment Analysis (DEA). The use of a graphical approach to identify the best compromises is not possible in the case presented here because four PMs have been included. This fact makes the use of the proposed approach completely necessary to solve the problem at hand.

Multiple quality criteria optimization in reactive in-mold coating (IMC) with a data envelopment analysis approach

Reactive in-mold coating (IMC) products have been used successfully for many years to improve the surface quality of Sheet Molding Compound (SMC) compression molded parts. IMC provides a smooth, sealed surface used as conductive or non-conductive primer for subsequent finished painting operations. The success of IMC for SMC parts has recently attracted the interest of thermoplastic injection molders. The potential environmental and economic benefits of using IMC as a primer and, in the ideal case, to replace painting completely are large. Most optimization studies in Reactive Polymer Processing involve a compromise between different performance measures. In most cases the controllable variables have a conflicting effect on the relevant performance measures. IMC is not the exception. These performance measures need to be balanced, each against the other, in order to obtain the best compromise. The ideal compromise will depend on the final part quality requirements. In this work, the use of Data Envelopment Analysis (DEA) is explored to identify the best compromises among several performance measures. We have selected two case studies to illustrate the use of this technique. In the first case, we apply DEA to select the locations for two IMC injection nozzles for a thermoplastic part to optimize two quality measures. In the second case, we study the simultaneous optimization of cycle time, surface quality, and dimensional consistency for SMC parts. The first case is aimed to demonstrate the application of DEA with a simple example; in fact, the best compromises in such example could have been identified graphically. The second case, however, provides an example where the multidimensionality of the problem makes the use of DEA critical to elicit a proper solution.

Determining the number of spares in an inventory/repair system which supports equipment with scheduled usage

Abstract We address the problem of finding the minimum number of spares for systems of equipment which are used on a scheduled, periodic basis. When spares are expensive it is important to have as few spares in the system as possible subject to managements specification of the required likelihood of having equipment available for use when needed. Several varieties of this problem are addressed in the literature, but all of them assume the equipment is needed or must be available all the time rather than according to a schedule. We present the base case of a series of stochastic models that can address several versions of this problem and which explicitly address the aspect of scheduled usage. Numerical examples based on NASAs space shuttle are presented. A comparison is made between the results of our model and an approach in the literature that is in common use. The comparison shows that the usual approach errors on the side of too many spares in the system.

A Dynamic program for designing factorial experiments with budget constraints

Abstract This paper develops a dynamic programming formulation of the factorial experimental design problem with budgetary constraints. The approach is valid for any number of factors if all possible interaction terms are assumed in the model. A specific structure involving no interaction terms (additive model) is also formulated. An example is presented along with some computational results.

Selection of cost-optimal 2k−p fractional factorials

Selecting an optimal 2k−pfractional factorial is structured as a mathematical programming problem. An algorithm is defined for the solution, and the case of additive costs is shown to have a known solution for resolution III designs.