SOLVING A SCHEDULING PROBLEM : BY RULE-BASED SYSTEM WITH DP AND LP ENGINES
スポンサーリンク
概要
- 論文の詳細を見る
Scheduling problems are both classic and new. They are classic since they were recognized many years ago as a challenge, especially in the field of manufacturing. They are new in that their environments have changed in two important aspects : (1)the newly emerging concept of flexible manufacturing systems (FMS)that require complex decision-making has introduced more complexity into scheduling problems, and (2) to cope with this complexity, it has become essential to use highly advanced computer systems A scheduling problem is characterized by two difficulties. One is combinatorial explosion : an n-machine, m-job problem has (m!)^n possible schedules, so that without elaborate and intelligent methods, a prohibitively large number of cases must be checked. The other is the diversity of conflicting constraints : a problem is usually constrained by due date, cost limits, production levels, machines, order characteristics, resources, and other factors. Scheduling problems have been extensively studied by a technique called Operations Research (OR). OR is an analytical method for obtaining an optimal solution by modelling. Recently, many complex problems have been investigated by using Artificial Intelligence (AI) techniques, since it is very difficult to address them by using analytical methods and conventional computer technologies. One important success in the field of AI is Expert Systems (ES), also known as Knowledge-Based Systems. ES exp1oit human experts' knowledge represented in the knowledge base. The purpose of the present paper is to propose a new approach to scheduling problems that uses a rule-based system with deterministic algorithms in an engine. In section 2, an actual scheduling problem for steel-making processes is described. In section 3, a basic AI approach and analytical solutions are presented. In section 4, a rule-based system with Dynamic Programming and Linear Programming engines is proposed.
- 一般社団法人情報処理学会の論文
- 1988-09-12
著者
-
Numao Masayuki
Ibm Research Tokyo Research Laboratory Tokyo
-
Morishita Shin'ichi
Ibm Research Tokyo Research Laboratory Tokyo
-
Tanaka Tsunehiko
IBM Research, Tokyo Research Laboratory, Tokyo
-
Tanaka Tsunehiko
Ibm Research Tokyo Research Laboratory Tokyo