Under the assumption that any process-failure mechanism follows an exponential model, have been studied the economic design of X^^-control charts and the determination of the sampling intervals for process control in many literatures. Recently, when the demand and production rates are given, how to determine the optimal production cycle and the economic production quantity has also been studied, so as to minimize the expected total cost per unit time including the set-up cost and the holding cost. In this paper, we propose a new integrated model of the process control and the inventory control. Concretely, under the assumption that the process-failure mechanism follows a Weibull distribution having an increasing failure rate, how to jointly determine the optimal X^^--control chart, sampling (inspection) interval, production cycle time and economic production quantity is discussed, so as to minimize the total expected cost per unit time including the process control cost, the set-up cost and the holding cost. Furthermore, the following two cases are compared and argued. One is where the process is kept running, and the other where the process is stopped, while searching for a possible assignable cause when the X^^--control chart shows an out-of-control state.
There have been a huge volume of researches concerning dispatching rules in job shop environments Strangely, studies all but one pay no attention to multiple identical jobs due to the existence of production orders (or batches) each of which calls for a fixed quantity, called an order size, of a specific part. Jobs arrive in the form of production orders in most, if not all, job shops. This paper presents simple processing-time-based dispatching rules, and shows experimentally based on simulation that the rules with order-size considerations perform consistently well for tardiness and utilization measures. Several alternative strategies for rule specifications are examined. The paper also considers finiteness of system entitles and of buffer sizes, and studies their effects on the performance of dispatching rules.
In JIT environment, the basic assumption is that every job should be completed before or just on the due time when the customer requires it and leave the shop simultaneously on the due time. If a job is completed earlier than its due time, then the shop holds the job and incurs its holding cost for earliness. When the single machine scheduling problem to minimize the total weighted earliness under the constraint of due times is considered, a given problem will not be always feasible due to the conditions that jobs are not permitted to be tardy and starting times for all jobs are non-negative. Such a case causes trouble in practice, and suitable actions will be taken. This paper addresses the scheduling problem to meet due times with reduction in processing time for such infeasible problems. A heuristic scheduling algorithm is developed to minimize the sum of the reduction cost in processing time and the holding cost for earliness. The computational experiments are also shown for the proposed algorithm.
This paper deals with the problem of evaluating two types of flexibility in job shop which has a bottleneck machine due to various factors. We compare machine flexibility and routing flexibility, especially for bottleneck which has various characteristics, in terms of manufacturing performance under various environments such as flow shop and random job shop. Based on the result, we present a basis of designing an efficient flexible manufacturing systems.
The main purpose of this paper is to build a business evaluation system based on the experts' knowledge and experiences through a neural network which realizes non-linear discrimination and generalization. In the modeling of business evaluation using business evaluation cases, the neural network is taught the input of 5 financial indices and the output of 4 clusters, whose classification is given by experts based on the financial indices. Furthermore, an enhanced method for learning algorithm is proposed in order to shorten the learning time to a larger extent than a conventional method. By its use in the discrimination of samples generated by an exclusive OR and a sine curve, its effectiveness is shown to shorten their learning cycles up to 70% and 50%, respectively. Lastly, we show the effectiveness of the enhanced neural network model for business evaluation to assess the corporations of industrial equipment using real data.
In a work which is mainly carried out by hand motions, there exists a series of transformation steps to convert raw materials into final products. In such work system, a series of workers' right and left hand motions are controlled by themselves. In addition, the workers also control functions and movements of tools through their hand motions. Furthermore, raw materials are transformed into final products through the process which is controlled by these hands and tools. Thus, a work system is comprised of various "control objects" and their "transformation steps", and they are playing their functions in the system in the manner of means-ends relationship. Examination of the adequacy of means for higher objectives becomes possible if this means-ends relationship can be systematically analyzed. This paper proposes a new mehod for work system analysis which describes the controlled objects as "handled resources" and their transformation steps from the viewpoint of means-ends relationship. Fundamental concepts for pursuing a simplified work system which suits the ultimate goal of obtaining final products are also investigated.