Shigeyuki Tomita

Design of flexible production system by means of tax-system-like algorithm (TA)

Abstract In most of design problems of modem production systems, it is always required to find the most appropriate configuration of composing unit that can achieve the required system performance with maximum efficiency. A new algorithm called tax-system-like algorithm (TA) has been developed to cope with this difficulty by taking autonomous agent system paradigm as its basis; it has also been developed by mimicking such ability of tax systems as can lead an activity level of a society to desired one by manipulating its tax rate. An appropriate configuration of an agents cluster that works at a desired utilization rate can be self-organized through this algorithm, where the number of agents is adjusted by the balance between rewards and tax payment. A computer based simulation system was also implemented and a series of experiments was carried out in order to evaluate the ability of this approach.

On an object-oriented modeling of supply chain and its operational strategy

Abstract In recent years, researchers have devoted a great deal of attention to supply chain management (SCM) and various production models have been proposed. In such chains, a lot of decisions have to be made to achieve the production objectives under a variety of uncertainties. In order to operate SCM efficiently, it is indispensable that some rational and sound operational strategy for realizing cooperative operation among such chains. A simulation-based approach is one of the most effective ways as a design method of a production model together with its strategy under uncertain conditions. In this research, an attempt was made both constructing such a environment by means of Object-Oriented approach and, utilizing such a environment, exploring the possibility of a new production model between a maker and suppliers where both “Build-To-Order” and “Build-To-Stock” are used to suit the customers needs. The validity of the model and the strategy proposed here was examined through computational experiments by using a simulator developed based on this approach.

Extraction of the Functional Structure Inhering in a Chemical Plant by Means of the Functional Model.

化学プラントのより知的な運転管理を実現ずるためには, 対象プラントに埋め込まれている諸機能を計算機に認識・理解させることが不可欠である.本研究では, 別報で提案した「機能」モデルに基づいて, 個別のプラントのPFDおよび物質の状態から対象プラントに内在する諸機能の階層構造を抽出し, その説明構造を自動に生成する方法を開発した.機能抽出は, まず「着目ユニット」の考え方に基づいて対象プラントに固有の「主機能」が抽出され, 次に機能間の階層関係に関する知識を用いて, プラントの機能の階層構造が生成される.このとき必要となる「最下位機能」の満たすべき構造的条件や各機能の意味に関する知識は明確なデータ構造で与えられており, これを用いて機能の階層構造についての説明構造が生成できるようになっている.