Knowledge that Transforms

Market globalization, higher requirements for improved quality, and tough, faster‐pace, price‐sensitive competition have led to two parallel and visible quality thrust: the Baldrige Award in the U.S. and, internationally, the ISO 9000 standards. The relationship between the Baldrige Award and ISO 9000 registration is widely confused. Two common misper‐ceptions stand out: (1) that they both cover the same requirements and (2) that they both address improvement, relying on high quality results, and thus, are both forms of recognition. Many have concluded that the Baldrige Award and ISO 9000 are equivalent and that companies should choose one or the other. These conclusions are incorrect. The Baldrige Award and ISO 9000 registration differ fundamentally in focus, purpose, and content. The focus of the Baldrige Award is on enhanced competitiveness. The Award Criteria reflect two key competitiveness thrusts: (1) delivery of ever‐improving value to customers and (2) improvement of overall operational performance. The Award's central purpose is educational‐to encourage sharing knowledge and experience of competitiveness and to drive this learning, creating an evolving fund of knowledge. By contrast, the focus of ISO 9000 registration is on conformity to practices specified in the registrant's own quality systems. Its central purpose is to enhance and facilitate trade. The Baldrige Award addresses competitiveness factors either not addressed in ISO 9000 registration or addressed differently. These factors include a customer and market focus, results orientation, continuous improvement, competitive comparisons, a tie to business strategy, cycle time and responsiveness, integration via analysis, public responsibility, human resource development, and information sharing. Overall, ISO 9000 registration covers less than 10% of the scope of the Baldrige Award Criteria and does not fully address any of the 28 Criteria items. As a result, the national drive to improve competitiveness could be diminished. Companies required to or electing to seek ISO 9000 registration are encouraged to integrate their conformity efforts with the Baldrige Award competitiveness improvement framework.

As manufacturing firms push to achieve shorter lead times and higher levels of customer service, the basic capability of underlying manufacturing processes must be reexamined. The capacity and operational variability of a process dictate a certain set of realistic performance goals. In this paper, we examine this fundamental relationship from an economic perspective using two levels of analysis. At the aggregate level, we model the manufacturing process as a single server queue and compare the traditional roles of marketing and manufacturing in setting performance and process design parameters. Insights gained at this level are incorporated into the analysis of a realistic multiserver, multistation manufacturing line. We develop an interaction decision tool to guide the selection of process and performance parameters in this more complex environment.

We have developed an improved implementation of the Wagner‐Whitin algorithm for economic lot‐sizing problems based on the planning‐horizon theorem and the Economic‐Part‐Period concept. For many rigorous test conditions, this algorithm is about twice as fast and requires only half the array storage capacity of the previously fastest algorithm. Its execution time is approximately linear in the number of periods in the planning‐horizon.

Originally conceived by Frank and Lillian Gilbreth, the “tabletop improvement experiments” have been used in Japan since 1925 to teach important principles of continuous improvement. The experiments, designed for classroom use, communicate their lessons in a striking and memorable way. The work‐related experiments categorize the sources of resistance to change and show how to neutralize them. The process‐related experiments sharpen understanding of where the biggest opportunities for process improvement usually lie. Surprisingly, the experiments are hardly known in the West. We describe all of them and document their history for the first time.

Recent theoretical work suggests that quality‐improvement activities can yield significant indirect effects through process improvements and reduced factory congestion and confusion, benefits that are overlooked or hidden in most management accounting or cost of quality systems. Using time series data from two consumer durables manufacturing plants, I estimate the indirect productivity gains from quality improvement. The evidence from the plants indicates that the indirect effects from improved quality are at least two to three times the direct benefits attributable to lower scrap, rework, and inventory holding costs. An important implication of these findings is that companies that justify investments and measure performance based only on the direct costs of poor quality will motivate managers to make suboptimal decisions regarding quality‐improvement activities.

This is a case study of workforce scheduling in the U.S. postal system. We use it to analyze the benefits of scheduling flexibility at postal distribution systems, which can come from several sources. We focus on the additional flexibility deriving from increasing the proportion of part‐time employees, as well as from increasing the cross‐training of part‐time employees. These two dimensions of scheduling flexibility are decision points of particular interest to the postal management. We used a large‐scale simulation model of the dynamic functioning of the postal distribution system in conjunction with a staff‐scheduling model to obtain insights into policy issues of interest. Our results show that the operating environment at a distribution station affects cost and customer service performance, and that gains from greater scheduling flexibility are situationally dependent. The benefit of cross‐training part‐time workers is modest over the range of levels considered realistic for distribution stations. Our overall recommendation is for postal managers to increase the proportion of part‐time employees, allowing more efficient matching of resources with the varying demands for mail delivery on different routes. For the range of conditions we modeled, cost reductions from more part‐time workers average over six percent when capacity utilization is low. Customer service improvements are even larger, and range between 20 and 43% when capacity utilization is high. Improvements at the upper end of the range are achieved when demand variability is also at its highest.

The focus of this work is on the effects of learning on economic production quantity in batch production systems. We assumed that both unit variable manufacturing time and setup time follow a learning curve. We modified the classical Economic Production Quantity model to incorporate these two types of learning phenomena. We also incorporated the forgetting effect in our model so that a fraction of the learning is lost between consecutive lots. We developed a dynamic program to obtain the optimal solution to the problem. We investigated the nonincreasing lot size property and used it to improve the efficiency of our dynamic program. We consider a special case of the model in which all lot sizes are assumed equal. After theoretical treatment, we carried out a computational study of the effect of assuming equal lot sizes on the optimal solutions. The results of our examples strongly indicate that the assumption of equal lot sizes not only simplifies the determination of the optimal solutions, but also provides close approximations to the optimal solutions.

Run‐length distributions for various statistical process‐control charts and techniques for computing them recently have been reported in the literature. The real advantages of knowing the run‐length distribution for a process‐control chart versus knowing only the associated average‐run length of the chart have not been exploited. Our purpose is to use knowledge of the run‐length distribution as an aid in deciding if an out‐of‐control signal is a true signal or merely a false alarm. The ability to distinguish between true and false signals is important, especially in operations where it is costly to investigate the causes of out‐of‐control conditions. Knowledge of the run‐length distribution allows us to compute likelihood ratios, which are simple to calculate and to interpret and which are used to determine the odds of obtaining an out‐of‐control signal at a particular run length when a shift in the process mean actually has occurred vis‐a‐vis no such shift. We extend our analysis in a Bayesian sense by incorporating prior information on the distribution of the shift size of the process mean, combined with the likelihood ratio obtained from the run‐length distribution, to determine if a shift larger than a critical size has occurred. We give examples for the Shewhart chart, the exponentially weighted moving‐average chart, and the special‐cause control chart for processes with autocorrelated observations. The examples show that the current recommended usage of the average‐run length alone as a guide for determining whether a signal is a false alarm or otherwise can be misleading. We also show that the performance of the traditional charts, in terms of their average‐run length, can be enhanced in many instances by using the likelihood‐ratio procedure.

An important issue in planning capacity expansion under uncertain demand is the effect of lead time on the timing of plant construction. Our model helps decide whether (1) plant construction is initiated after a certain deficit is accumulated, or (2) plant construction is initiated ahead of demand when a certain capacity surplus is reached. In addition to our analytical results, we present computational results to show that it is economically attractive to delay plant construction beyond the time when existing excess capacity becomes fully absorbed, with relatively short construction lead time.

We address the issue of performance analysis of fabrication/assembly (F/A) systems, which are systems that first fabricate components and then join the components and subassemblies into a product. Here we consider an F/A system consisting of a single assembly station with input from K fabrication stations. We assume that the system uses a Kanban control mechanism with a fixed number of kanbans circulating between each input station and the assembly station. Even with Markovian assumptions, computing an exact solution for the performance evaluation of such systems becomes intractable due to an explosion in the state‐space. We develop computationally efficient algorithms to approximate the throughput and mean queue lengths. The accuracy of the approximations is studied by comparison to exact results ( K = 2) and to simulations ( K > 2). Part II of this paper demonstrates how these models can be used as building blocks to evaluate more complex F/A systems with multiple levels of assembly stations.