Assume that the quality characteristics obey a normal distribution. Recently, Kanagawa et al. (1993) proposed a (x^^-, s) simultaneous control chart based on the estimator of Kullback-Leibler information. Further, based on the results derived by Arizono et al. (1995), we can evaluate the power characteristics for out-of-control state using the simultaneous control chart. Then, the (x^^-, s) simultaneous control chart is designed to conform with the probability limit method and enables the user to monitor both changes in mean and variance simultaneously. By the way, Daudin (1992) proposed a double-sampling x^^- control chart, which is the counterpart to double-sampling plans. However, in a physical environment, it is usual to jointly employ the x^^- chart for monitoring the shift in process mean and the R chart for monitoring the change in process variance. Then, the application of the double-sampling procedure to the x^^--R chart is somewhat complicated as many design parameters and compound judgements are required. On the other hand, the (x^^-, s) simultaneous control chart for monitoring both changes in mean and variance simultaneously has less design parameters and simple judgement in comparison with the x^^--R control chart. In this paper, the double-sampling procedure is applied to the (x^^-, s) simultaneous control chart, then a double-sampling (x^^-, s) simultaneous control chart is proposed. Furthermore, the performance of the double-sampling (x^^-, s) simultaneous control chart is compared with the performance of the standard (x^^-, s) simultaneous control chart.

View full abstract

This paper proposes an image processing method for arraying and inspecting flexible belt-shaped objects. First, a method using a labeling procedure is presented to distinct folding patterns of belt-shaped objects. Next, a grasping position on a object is determined considering the width of the robot hand and folding pattern of the object. The grasping position is used by the robot hand, which moves so as clearly to untangle the object. In addition, the shape of the object is adjusted so that inspection points of the object can be roughly overlapped to the same points of its CAD image. Then, an inspection procedure is presented to detect identical defects of the object by comparing the contour lines of the object and its CAD image. Finally, validity of the proposed method is evaluated through an actual application for arraying and inspecting a belt-shaped rubber object.

View full abstract

The objective of this study is to investigate empirically the two different approaches to the modeling of discriminate bankruptcy. We present a discriminate model of bankruptcy considering the equity and debt for a corporation's decision-making needs. Though the daily equity values are considered as a corporate evaluation by investors in the stock market, we can estimate the daily equity values of a corporation using daily stock prices by a beta estimate technique. On the other hand, it is difficult to estimate the daily debt values because a corporation announces the annual debt values only at the time of settlement of accounts. But we estimate the daily debt values using the Black-Scholes Model, which is used in the field of finance to evaluate option prices. As a result, we can discriminate corporate bankruptcy daily using our present model, which is constructed using by implementing daily equity and debt values. We can Show the validity of our model the model using real-world data.

View full abstract

A consecutive-k-out-of-n : F system is an ordered sequence of n components. This system fails if, and only if, at least the k consecutive components fail. The optimal assignment problem in linear consecutive-k-out-of-n : F system is to determine which assignment (optimal assignment) has the largest system reliability in all the assignments of components when all components don't necessarily have the same failure probability. In this paper, we propose two kinds of GA. The first GA considers all of the assignments of components and transfers them to chromosomes using ordinal representation directly. The second GA considers only the assignments of components holding necessary conditions for optimal assignment and transfers them using a revised ordinal representation. Some numerical experiments will show that the second GA is most useful of the two GAs and another approximation method.

View full abstract

This multiple autoregression model, named the Synthesis Forecasting Model, is constructed with two models : multiple regression and autoregression. The autoregression model is only based on numerical data ; However, in order to improve forecasting precision, it is necessary to be considered qualitative demand factor, such as customer satisfaction, market environment, and factory condition, affecting total demand. the Synthesis Forecasting Model can represent these factor weights and forecast as time series by synthesizing with autoregression, This proposed model is evaluated by coefficient of determination, f-test, AIC and residual test, and it is applied to the forecasting of sales quantities of MD radio casette player of an audio equipment maker. As a result of this application, it is clarified that the proposed model is an improvement over regular autoregression models and that Availability for actual data can be shown.

View full abstract

Past studies on project team productivity improvement were based on indirect approaches because productivity measurement indicators could not be applied to such studies. This paper proposes a model for productivity improvement in project team operations by applying the productivity measurement indicator proposed by Nakata et al. The authors suggested two hypotheses concerning factors affecting productivity in project team operations. First, from the viewpoint of the management process, the main factor is team activity planning. Second, as the project moves forward and importance shifts from theoretical research to implementation, the priority of team activity planning shifts from theoretical research to implementation. The authors investigated eighteen project teams whose goal was to reform the head office division or operation division in a certain enterprise at two points in time (the third month from the start of activity and the ninth month), and studied the productivity factors in project team activities and the relations between these factors and productivity in order to investigate the validity of the above two hypotheses. Findings showed that team activity planning accounted for sixty percent to eighty percent among the factors contributing to productivity increases. In the third-month investigation, team activity planning that focused mainly on theoretical research was the most important factor affecting productivity, and in the ninth-month investigation, team activity planning that focused mainly on implementation was most important. These findings verified that there was a difference in priorities of team activity planning between the two points.

View full abstract

The recent expansion of the network-computing environment in companies may affect their business styles, organization forms and workers' ways of thinking. To keep good balance between the technical system (information technology) and the social system (people, organization), we have to take users' viewpoints into consideration. We took a survey using questionnaires of three types for the system department, personnel department and employees in ten companies. Our analysis covers the utilization form of groupware/intranet and its affect on the change of jobs and the workers' ways of thinking. Also our research covers the purpose and effect of the use of groupware/intranet and the change of organization resulting from the use of groupware/intranet. Finally we propose a model of change by using information technologies (groupware/intranet).

View full abstract

This paper considers minimizing makespan in a machine-fixed, machining-assembly flowshop scheduling problem (MAFS) with M-parallel machines at the machining stage and one robot at the assembly stage. M-parts of each job are machined on prespecified machines at the machining stage and these parts are assembled at the assembly stage after all component parts for each job have been machined. Since this problem has been proved strongly NP-complete, we propose a branch and bound based heuristic algorithm, called the "squeezing branch and bound (squeezing B&B)" algorithm which searches for a prespecified number of promising nodes in parallel at each branching level and terminates the search without backtracking. For increasing the effectiveness of the squeezing B&B, a new lower bound is proposed and a dominance rule is incorporated into the branching process of the proposed method. The results of numerical experiments show that the proposed method can efficiently obtain an optimal (or best) schedule for almost all instances of the machine-fixed MAFS with two, three and five machines at the machining stage.

View full abstract

In this paper, we propose a multi-view monitoring system which displays a single object in different shots on various display devices taken by cameras on different sides in order to position a screwdriver at a screw, using a disassembling tele-operation system which is operated by remote manupulation for parts recycling in the disassembling process. It is assumed that the positioning accuracy for the screwdriver at a screw is influenced by such factors as the difference in monitoring systems, and the angle between the cameras' optical axes and screws' image size per pixel (resolution) in each display device. It makes clear by an analysis of variance that the positioning accuracy of depth and distance using the proposed monitoring system is higher than that of the stereoscopic vision system using specific glasses and display devices. And as the resolution is determined from 0.76 to 1.41mm/pixel under the conditions of an angle between cameras' optical axes from π/4 to π/2,the screwdiver can be set at the center of a screw within the acceptable circle of 1.0mm. Then as the resolution ratio between two monitors is given from 1 : 1 to 3 : 1 and the angle between cameras' optical axes is determined as π/2,the screwdirver can be set within the acceptable circle of 1.0mm under the condition that one monitor's resolution is set from 1.41mm/pixel to 1.85mm/pixel and the other's is 1.41mm/pixel to 0.77mm/pixel.

View full abstract

Cellular manufacturing is an application of group technology to manufacturing, and is a well-known strategy for reducing lead times in batch production systems. In essence, most of the group technology methods related to cell formation require crisp data for parts' and machines' attributes and hence dichotomously force parts and machines to fall into the classification scheme. However, in many situations, the parts' and machines' attributes are too obscure and are too complex to be susceptible to analysis by traditional methods of group technology. In such cases, this information will be described well using phrases in a natural language. Recently, Gill, et al. have proposed an approach, based on fuzzy linguistics, to deal with data quantification and similarity measure for part-family formation problems which are traditionally addressed through conventional binary coding structures. However, they only proposed the part-part similarity measure. In this paper, we newly add a method for making the part-family formation, part-machine similarity measure, and clustering method to allot the machines to the part-families. In addition, a performance measure and algorithm for finding the best cell formation are also proposed.

View full abstract

In this paper, a Markov analysis is applied for evaluation of differences between communication in teleconferences and face-to-face conferences. A discernment problem, which is that speakers are unable to discern the nonverbal cues, such as body language or eye contact, of their conversation partners in teleconferences is assessed with an experiment. In the experiment, two groups consisting of three subjects speak in a usual teleconference, improved teleconferences concerning discernment and a face-to-face conference, and the sequence of utterances which indicates who spoke after whom is obtained and coded. The sequence is regarded as a stationary Markov chain, and is tested for whether it is a first-order Markov process, a second-order one, or a third-order one. Moreover, mean first passage time, mean recurrence time and mean visiting time, which are typical quantities used in a Markov analysis, are calculated. These quantities showed that the conversation structure in the improved teleconferences concerning discernment was closer to that in a face-to-face conference than the usual teleconference.

View full abstract

The improvement of flexibility has become increasingly important as a method to achieve competitive advantage in manufacturing. In order to improve flexibility, management have to answer the following basic questions : What types of flexibility are required under environmental uncertainty? What is the impact of different types of flexibility on productivity, quality and competitive position? How can flexibility be measured? What are the sources of flexibility? How can a firm implement different types of flexibility? et at. This paper attempts to answer these questions by providing a conceptual framework where manufacturing strategies, type of flexibility, environmental uncertainty, flexibility dimensions and flexibility sources are set as variables. This framework helps to identify certain assumptions of theoretical studies, which need to be challenged. A general procedure is also presented for a firm to identify the series of managerial actions necessary to develop the action plans for flexibility improvement which will best contribute to its competitiveness. There are three main parts to this article. Section 2 provides a literature review and agenda for flexibility in the production system. Section 3 presents a conceptual framework for analyzing the strategic value of different types of flexibility. Section 4 provides a method for designing a flexible manufacturing system.

View full abstract

This paper develops an algorithm for due date assignment in a dynamic job shop. The objective is to provide an earlier due date for a newly ordered job to a customer, based on a feasible schedule for the already scheduled jobs and the newly ordered job. The proposed algorithm is based on a certain static job shop scheduling to meet due dates, and determines a new schedule under due date constraints for the already scheduled jobs so as to approximately assign the earliest due date for the newly ordered job. Thus, the producing agency proposes an earlier due date and then the producer and customer agree on a final due date.

View full abstract