1999 Fiscal Year Final Research Report Summary
Research on Systems Architecture to Support Technology Development for Production Scheduling Based on Case Database.
| Project/Area Number |
09680412
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
社会システム工学
|
|---|
| Research Institution | HOKKAIDO UNIVERSITY |
|---|
Principal Investigator |
SEKIGUCHI Yasuki Hokkaido Univ., Fac. of Econ., Pro., 経済学部, 教授 (00091471)
|
|---|
| Project Period (FY) |
1997 – 1999
|
| Keywords | decision support system / case base / production scheduling / modeling / problem definition / model / data independence / model / solver independence |
|---|
| Research Abstract |
(1) A new information model, Generic Entity-Relationship Model (GERM), and a new method for problem definition based on GERM were proposed. GERM is applicable for simulation as well as optimization problems. Architecture for a case database of problem definitions was proposed. A prototypical system of the case database based on the architecture was developed and effectiveness of the architecture was demonstrated. The case database can support users in describing real problems prior to mathematical formulation.
(2) A mathematical model to a GERM-based problem definition is defined as a pure mathematical object. The mathematical model is connected with the respective problem definition through information called "unification condition." This technique provides a method to describe many-to-many correspondence between mathematical model and problem definition. Based on the technique a case database to support mathematical formulation for problem definition can be developed and is convenient
… More
ly used for reusing mathematical models. Model/data independence is realized in the case database, because GERM-based problem definition is independent of numerical data.
(3) We distinguish "user defined model" that is a mathematical model developed by a user from "standard model" used by a developer of a solver. They are connected to each other through information called "connection information." This is our method for model/solver independence. A solver to a standard model is invoked from the user-defined model connected to it. This improves reusability of solvers. Besides, the user-defined model is connected to a problem definition (cf. (2)). As a result, a user can invoke a solver from a problem definition with numerical data.
(4) Experience on the prototypical model taught us that it is not easy for users to understand context of a GERM-based problem definition. This seems caused partly by the fact that object-types are fundamental elements of GERM. We proposed a method for inputting an object-type through input of its instances. Less
|
Research Products
(4 results)
