دانلود رایگان ترجمه مقاله روش جدید برای تشخیص تنگناها (نشریه IEEE 2008)

یک روش جدید برای آشکارسازی تنگناها

A NEW METHOD FOR BOTTLENECK DETECTION

Abstract

This paper presents a new method to identify and rank the bottlenecks in a manufacturing system. The proposed method is based on performance related data that are easy to capture, offers low computational burden and less prone to be in error due to its simplicity. The proposed method analyzes inter-departure time from different machines to identify and rank the bottlenecks. In a follow-up paper the authors plan to present a method to allocate loss of production to different machines based on analysis of inter-departure time. This paper also proposes a set of rules that may be used to improve data integrity. The proposed method may be used to analyze both steady state as well as non-steady state data and can be extended easily to analysis of a job shop.

1 Introduction

Detecting bottlenecks in a production line is not a trivial task. The study of the problem comes under the heading of Optimum Production Technology, lately labeled as Theory of Constraints. The problem of detecting the bottleneck is further aggravated by the fact that the manufacturing lines are inherently variable. The cause of this variation might be due to random events (machine failure) or long term changes in the system (seasonal variation of demand, new product launch, machine load changes). A production system may have a dominant bottleneck which will appear as the bottleneck most often during analysis or it may have momentary bottlenecks that keep shifting in the system with time. Shifting of bottleneck happens due to random event such as machine failure. Literature on bottleneck analysis refers to both momentary and dominant (average) bottleneck. There are various methods available for detecting momentary and average bottlenecks. A partial list includes the following:

• Shifting bottleneck detection based on the duration of machine being active without interruption,

• Bottleneck detection based on utilization of machines, the machine with the highest utilization is considered bottleneck,

• The machine with the longest average up-stream queue length is considered bottleneck,

• Analytical approach based on estimation of the blockage and starvation probabilities of a machine,

• A graph theoretic approach to detect bottleneck in order to optimize the scheduling in a shifting bottleneck procedure.