This paper discusses how to control the process, whereby samples from the process are respectively classified into several categories by inspection. So far, when there are two categories. p chart or pn chart has been used, and when there are more categories, Chi-Square charts [1] Duncan proposed has been used. Recently, Tamaki et al. [6] proposed an advanced control chart based on Duncan's chart. These control charts are all based on the definition that the state of statistical control is the non-existence of variation between subgroups, and are designed by binomial distribution or multinomial distribution. However, a new control chart which considers variation between subgroups is required for the process control at the field in which many new products appear one after another in short cycle. This paper proposes a new control chart based on categorical data distributed as compound distribution and studies the level of significance of this chart.

View full abstract

Recently, Kimura et al. proposed a demand forecasting model for the purpose of predicting the demand of a directly succeeding period based on the past demand series using layered neural networks. In this paper, as an extension of their model, we consider a demand forecasting model for the purpose of predicting simultaneously the demands of several succeeding periods. Then, we propose a predicting method via layered neural networks based on differences of demands instead of demands in their model. Moreover, the effect of using the variable slope method, which the synaptic interconnection strength and the characteristic function of units are simultaneously renewed, is also investigated.

View full abstract

Selection of the variables and nonlinear terms of polinomials for approximating nonlinear multi-dimensional mappings has many difficulties, since it is a problem of a kind of combinatorial optimization. The group method of data handling (GMDH) is an algorithm for the nonlinear model identification from collected regression-type data. But, it needs the calculation of many combinations of polinomials which are called as the partial descriptions. The combinatorial algorithms are computationally intensive. This paper proposes a simple method of model identification using the fuzzy ID 3 and Bernstein polinomial, in which the B-spline is regarded as the membership function of fuzzy set. From a practical view point, an algorithm with less computational cost for selecting the terms of polinomial is discussed as well as links to rule extraction methds from numerical data. Since the obtained polinomial consists of the tensor product of B-splines, by regarding B-splines as the menber-ship functions of fuzzy sets, it can be represented as the fuzzy if-then rule type knowledges and fuzzy decision trees.

View full abstract

In this paper, we propose a method to design an assembly line to correspond flexibly to the global assembly style based on MRP. The precedence relation is separated into vertical and horizontal ones from the Bill of Material (BOM in MRP). The criterion for ordering the part and unit is introduced to determine the order of parts which is not determined by the precedence relations. The assembly line is constructed by the straight line to assemble the unit and by the parallel line to the subunit according to the BOM. In straight line or parallel line, the assembly station is constructed from the several part supplying stations by the criterion that the assembly time is less than the pitch time and the area covering by a man or a robot is restricted.

View full abstract

This paper, taking the difference of assembly time of each product into account, proposes a new heuristic method based on the goal chasing method for sequencing various-assembly-time products in a single work station. In addition, the effect of neglecting the difference of assembly time of each product is discussed through comparing a number of results obtained by the goal chasing method and our proposed method. A large number of better results show the effectiveness of our proposed method.

View full abstract

This paper presents a design method for mixed product lines constructed of turret robots and then the validity of the method is proved. This method selects a kind of robot for each work station and kinds of tools to be set for the robot and assigns elemental works to each robot from the viewpoint of minimizing the number of work stations in the line to execute desired production plans. This method is a branch-and-bound method. Its main use is to assess practical design methods and solve small design problems.

View full abstract

A supplier Kanban system controls the orders to a vender in a Just-In-Time (JIT) production system. In our previous paper, we have proposed an algorithm for determining an optimal number of supplier Kanbans and an optimal withdrawal cycle that minimize an expected average cost per period under stochastic demand. This paper deals with stochastic properties of the supplier Kanban system and its optimality. First, we show behavior of the backlogged demand using the stochastic ordering and convex ordering in the case where the number of supplier Kanbans, production capacity and demand distribution change, respectively. Next, a problem of deter-mining an optimal ordering policy without using supplier Kanbans that minimize the expected average cost per period is formulated into a Markov decision process (MDP). Finally, an optimality of the supplier Kanban system with the optimal number of Kanbans is discussed through numerical comparisons between the supplier Kanban policies and the optimal ordering policies computed by the MDP.

View full abstract

The objective of this paper is to formulate the optimal control system for the linear system with colored noise and to apply to the periodical ordering system for the push-type production and inventory systems, which provides the possible procedure to control both the variances of order quantity and inventory. This system realizes the optimization of the criterion of the weighted sum of the variances of production and inventory. For the extension to the solution of multi-stage process and of the case the demand is not stationary, it is necessary to formulate the control system by the optimal control theory. However, the optimal regulator control which is the conventional optimal control theory can be applied directly to the system with colored noise. In this paper, we propose the new formulation of this problem and the application to the production and inventory control system.

View full abstract

Flexible Forging Machine (FFM) is one of the most important machines in general flexible manufacturing system. The scheduling problem of FFM is to reduce or preferably eliminate changeover cost, and is an NP-hard combinatorial optimization problem. The Genetic Algorithm (GA) is known as one of the most effective heuristic search algorithms, and suitable for solving this kind of problem. However, applying GA to this problem, we frequently obtain a local optimal solution rather than a best approximate solution. The goal of this paper is to solve the above mentioned problem about local optimal solutions by introducing a measure of diversity of populations using the concept of information entropy. Thus, we can obtain a best approximate solution of the scheduling problem of FFM by using this advanced GA.

View full abstract

This paper develops a study of Flexible Manufacturing Systems with Dynamic Routing. First, the performance of the system is reevaluated by using the theory of closed queueing networks, and compared with the corresponding simulation model. Next, the behavior of the system is compared with the case of fixed routing, and then, it is found that it is an increasing concave function of finite local buffers, and so on. Further, the two configuration problems that considered these properties are formulated. Finally, optimization algorithms using implicit enumeration method are proposed, and a numerical example is given.

View full abstract

A production system to process multi-item products by using a single machine by turns to meet time-varying stochastic demand in finite horizon consisting of discrete periods is considered and a new optimization model and its calculation procedure are proposed. The problem is to determine the item and the production amount to minimize the expectation of the total of production costs, inventory-holding costs, shortage costs and set-up costs. The problem can be formulated as a Markov decision process, but it is not practical because its dimension becomes too large. Therefore an eclectic model is built, where the item is treated as a variable to be determined at the beginning of the planning horizon and the production amount is determined as a policy. Two propositions to make the characteristics of the problem clear are introduced and the method for obtaining the lower bound to the partial problem is proposed based on the propositions. A calculation procedure using the branch and bound method is developed. Meaning and validity of the proposed model are clarified by the numerical example.

View full abstract