In conventional data envelopment analysis (DEA), there are cases where the decision making units (DMUs) that cannot be considered efficient according to past experience are determined as DEA-efficient. This paper proposes a DEA using virtual DMU as intermediates (DEA/Virtual DMU Analysis) to conduct evaluations in accordance with the evaluator's senses. Under the proposed method, the evaluator's a priori information is materialized as virtual DMUs and added to the real DMU set for the purpose of analysis. Accordingly, this method enables the evaluator to determine whether the DMUs satisfy the activity level presupposed by the evaluator, thereby making it possible to reflect the results on DEA-efficiency values. This paper also introduces an application of the proposed method to business analysis of Japan's automobile manufacturers in order to demonstrate the effectiveness of the method.

View full abstract

Considerations of researchers regarding new product development include the planning of product strategy, formation of product theme, and cost management through target costs and so on. These factors have been studied independently. Thus, there has been no consideration of the relationship among these factors, as well as the relationship between these factors and profitability. The aim of this paper is to determine the relationship among these factors by causal analysis through three structural equation models and to confirm that companies that maintain a successful relationship among factors may attain good profitability.

View full abstract

This paper discusses an efficient algorithm for optimizing the operation of an independent power plant by extraction steam turbines often used in chemical plants. This problem is formulated as a nonconvex programming problem for which it is very difficult to obtain an exact optimal solution. The algorithm, which uses a linear programming relaxation and local search method, finds a nearly optimal solution very fast. We believe that this algorithm can be used for the real-time control of a power plant.

View full abstract

As an order release system for multi-stage production inventory systems, the Kanban system has been developed, and its alternatives, the concurrent ordering system and constant work-in-process (CONWIP) system, were also proposed. In the Kanban system, an order is released not immediately after a demand arrives, but after the production or transportation for the demand is available. On the other hand, in the concurrent ordering system, the order is released to all the stages immediately after a demand arrives. In the CONWIP system, an order is released only to the first stage when a demand is satisfied by the product, and the other stages are controlled by "Push" concept. The difference of the ordering mechanism leads to characterizing the performance of the ordering systems. In this paper, the difference in performance among the Kanban system, concurrent ordering system, and CONWIP system is investigated. For the mathematical model of each ordering system constructed, the structural properties are clarified and quantitative performances are analyzed by simulation experiments. As the buffer size affects the performance, the algorithm for allocating buffers is shown for each ordering system.

View full abstract

This paper presents a new approach to the coordinate type facilities layout problem. The site and each facility location are represented in a coordinate system. The objective is to minimize the total cost incluing the material handling cost and site areas cost, and maximize a closeness rating measure. We present "locatable edges" for the facility arrangement in order to restrain the search space. The results indicate that the model does an excellent job for the facilities layout problem with unequal areas as well as with equal areas, and that the ability to handle large scale problems is also improved.

View full abstract

Cumulative curve is an analytic technique that is used for the improvement of production processes and planning in manufacturing. For example, the cumulative curve can be expressed by accumulated generating operation of input and output data in a manufacturing system. Here, the lead time is shown as a horizontal interval between the input and output curve. The inventory is shown as a vertical interval between the input and output curve. We have already proposed that the cumulative curve can be expressed as GM(1,1) and GM(1,2) models. In this paper, we suggest a method to solve the lead time and inventory between arbitrary processes with GM(1,1) and GM(1,2) models. The equation to compute the lead time and inventory can be proved mathematically with the GM(1,1) and GM(1,2) models. We prove this with numerical experiments. As a result, lead times and inventory can be solved with coefficients of GM(1,1) and GM(1,2) models of cumulative curve and the initial value of time series data. Cumulative curve can be expressed as a transfer function. The models discussed are illustrated with numerical examples.

View full abstract

This paper considers scheduling in a multi-stage flow shop where the final stage is the "assembly station" as studied by M.Yokoyama in 1996 and 1997. Several "parts" are processed in the earlier stages and assembled into a product. We assume a so-called lot-streaming environment which allows partitioning of a lot into a number of equally-sized "sublots", and we try to construct a schedule which smoothes out the completion times of individual products. An enumerative algorithm is presented to solve the problem, and is evaluated via computational experiments. The proposed algorithm can solve problems with less than 8 different part types. The original algorithm works well when the lot can be divided into equally-sized sublots in several different ways; that is, considering the fact that the algorithm only works well when lot size D has a reasonable number of divisors, two variants of the approach are also presented which are applicable to cases where lot size D has only a few divisors. The second variant is shown experimentally to produce good results within reasonable computing time.

View full abstract

This paper outlines an estimation method of structural failure probabilities based on simulation within a limited sampling region. The basic random variables are assumed to be independent standard normal variables and the reliability index β to be known. To estimate such a small value of failure probability through the crude Monte Carlo simulation, a very large amount of samples as well as long computer processing time are usually required. To save computational efforts, a new sampling method is proposed to generate samples restricted outside the β-sphere, whose radius is the reliability index β, and within an effective distance range which is defined as a range to have effective probability values. In the proposed simulation method, a truncated chi-square distribution is introduced. Each sample within the limited sampling region is produced by combining a random radius generated from the truncated chi-square probability density function with a random directional vector generated from standard normal samples. Numerical examples are presented to show the validity of the proposed method.

View full abstract

Products can be classified into functional and innovative types according to the demand pattern. The lifecycle of the functional type is relatively long and daily demand is relatively stable. A traditional supply chain model, called efficient supply chain, can be applied to this type of product with cheaper management costs. On the other hand, the lifecycle of the innovative type is short, and the demand grows rapidly to reach a peak and then declines rapidly. If we apply the efficient supply chain to this innovative type of product, the model may result in a large number of shortages in the growth period, and high level of dead stock at the end of the product lifecycle. Our purpose of this research is to construct a quantitative supply chain model, called "responsive supply chain", which can react sensitively to the market changes, and therefore capture the opportunity of high profit exactly in the growth period with sufficient supply of product and reduce hackneyed costs at the end of the product lifecycle by adjusting the product supply quantity after peak demand is reached. In a particular numerical examination, applying the responsive supply chain model to an innovative-type products, we obtained large percentage reductions in average shortage and average inventory simultaneously, compared to applying the effective supply chain model to innovative-type products.

View full abstract

In local areas, there are many small companies that can make and sell goods. They supply goods marked by their strong individuality. But, their stores are few, and their profit is greatly affected by customer needs. On the other hand, convenience stores have many goods categories and they are dotted about the area. Therefore, convenience store becomes conspicuous recently. In previous study, mutual supply system was proposed with consideration of supplying many goods categories and goods marked by their strong individuality. This system's objective is to provide sales opportunities for small companies. In this paper, we make a comparison between route delivery and distribution center delivery for logistics of the mutual supply system. We show the formulation of distribution costs and also consider delivery truck capacity.

View full abstract

For manufacturing companies, it has become important to decide which technology is a future leveraged technology and how much to invest in R & D. That is, the planning skill for strategic technology is a key factor for permanence. However, companies don't create only technology strategies but also technology stock. A systematical methodology for recognizing technology stock is required at present. Technology recognition is a group decision problem over plural divisions. In this study, we propose a methodology for extracting and assessing technologies and develop an assessment system for plural decision-makers. Concretely, we focus on technologies that are tangible for products, assess the technologies extracted from production processes and quantify them by "total technology assessment." Utilizing this system, it is possible to recognize a core technology in a company and thus support future product development.

View full abstract

In this study, three methods for evaluating the constants a, b and c are proposed using the function y(x) = ae^<-bx> + c, which is suited to realities in place of the exponential function y(x) = ae^<-bx> being used besides the power function, as a learning curve. In the first method, b is evaluated by linear regression analysis and a and c are evaluated by linear regression analysis using the obtained b. In the second method, c is evaluated by linear regression analysis and a and b are evaluated by power regression analysis using the obtained c. In the third method, a, b and c, minimizing the residual sum of squares, are evaluated using the simplex method. Then the initial values of a, b and c are based on the values obtained in the first or second method. Further, the effectiveness of the proposed methods is confirmed using numerical examples.

View full abstract

Determining cutting conditions (tool replacement time, cutting speed, and so on) is a critical component for successfully meeting production schedules and economic demands. It is well known that tool life is probabilistic in nature. This paper presents a generalized machining cost model based on a stochastic model of tool life under two failure modes (i.e., fracture or wear) for determining economic tool replacement and cutting speed. An interruption of machining due to in-process sudden tool fracture or excessive tool wear is detrimental in automated machining systems. Not only does it halt production, but it also affects the quality of the machined surface. The workpiece in-process becomes defective. The interruption of the cut results in the loss of valuable time and an increase in production cost. The optimal tool replacement time and cutting speed are prerequisites for the design of an efficient and economical machining system. The losses incurred by tool fracture are different from those incurred by excessive tool wear. This paper formulates a new machining cost model based on the losses arising from tool fracture and in-process excessive tool wear. The distribution of tool life caused by tool fracture or excessive tool wear is given by the competing risk model under the dependent competing causes of failure. Optimal tool replacement time and cutting speed are obtained with the minimum production cost criterion.

View full abstract

Under the just-in-time (JIT) pull system, the sequencing of cars requires the satisfaction of two main goals : (1) keeping a constant rate of usage of parts, and (2) smoothing production load at workstations to avoid line stoppages. By using a practical observation, we focus our study on the second goal, where we investigated the effectiveness of a re-sequencing rule-based approach commonly used in the automotive industry in comparison with a more general time-based approach. We designed a simulation experiment based on practical situations in a final assembly line and found that the rule-based method presents a better achievement in terms of production load smoothing.

View full abstract

In this paper, we propose the estimation and reduction methods for manufacturing costs based on a proposed work structure diagram. Firstly, we construct a work structure diagram in which activities corresponding to transportation during processes, operation/assembly, inventory, and logistics both domestically and in foreign countries are added to the structure diagram in MRP. Secondly, we propose an estimation method for manufacturing costs based on the work structure diagram. In the breakdown method, manufacturing costs are calculated from top to bottom in the work structure diagram, and in the bottom-up method, cost is calculated from bottom to top. Manufacturing costs are estimated based on not only cost factors but also fixed and variable costs. Thirdly, we propose a reduction method for manufacturing costs in the breakdown method, bottom-up method and imported method based on unit selection criteria (maximum cost, mean cost, fixed and variable costs) and foreign manufacturing criteria (inferior rate of unit and exchange rate). In the breakdown and bottom-up methods, variable or fixed costs are reduced to attain profits under the condition that the variable cost rate of a process is given and the break-even cost is calculated to the given variable cost rate. In the imported method, parts/units are imported from a foreign country where the import costs are minimal.

View full abstract

Statistical model selection is one of the most important problems in statistics, and many works have left essential results. The conventional information criteria for model selection, such as the Akaike information criterion (AIC), the Bayesian information criterion (BIC), and minimum description length (MDL) were derived from different viewpoints. Many other model selection criteria have also been reported from various viewpoints. On the other hand, if we specify the model class and assume prior probabilities, then we can acquire Bayes optimal model selection for a finite number of samples based on Bayes decision theory. Furthermore, we can assume the various loss function adapting the purpose of model selection for practical cases. In this paper, we analyze the asymptotic properties of stasistical model selection based on Bayes decision theory. At first, we formulate Bayes optimal solution based on Bayes decision theory. In this formulation, we introduce the general loss function for practical problems. Moreover, we analyze the upper limits of the error rate of the model selection.

View full abstract