Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/85262
Title: A heteromorphic paradigm for networked production system
Authors: Wang, Qing
Degree: Ph.D.
Issue Date: 2004
Abstract: The globalization of manufacturing has brought extensive changes to the Production Management Systems (PMS) of enterprises. To stay competitive, enlightened enterprises have strived to achieve greater cooperation and collaboration among business partners in an approach called supply chain network or the virtual enterprise. The new role of the Distributed PMS (DPMS) is investigated from four points of view: the architecture of the DPMS, the development model, the implementation platform and operations management. The work reported in this thesis defines a new architecture for DPMS called the Networked Joint-Production System (NJPS). NJPS is derived from a new concept of Joint Production, which can be viewed as a counterpart of what the Joint Venture is to financial activities. It aims to facilitate autonomus collaboration between manufacturing partners at the tactical level of production management by the use of intelligent services. In order to develop NJPS a new development model called the Hierarchical Multi-View Model (HMVM) has been designed. HMVM describes the NJPS from multiple view points in a heteromorphic paradigm so as to support agent-based applications. The HMVM is platform-independent and may be customized for any system development platform. An XML-based (eXtensible Markup Language) General Reconfigurable Agent Model (GRAM) is proposed to support the implementation of NJPS. GRAM aims to facilitate the the implementation of the applications of an Agent-Based System (ABS) by designing a plug-in function repository, a set of controllable XML based communication and a multi-thread mechanisms. It may be used as a general platform for intelligent applications and for integration with a conventional legacy system. Various intelligent applications may be implemented with the ABS based on the proposed GRAM in order to support decisions in the operations management of NJPS. An Order Dispatching Problem (ODP) is selected as a key point in the order management. Two mathematical models have been designed to solve the ODP. The usage rate of parts inventory and production capacity have been considered in the two models, which are important factors in production management. Two optimization methods, a Neural Network (NN) incorporated with a heuristic mechanism and a Genetic Algorithm (GA) with embedded Pattern-based Evolutionary Mechanism (PEM), have been used to find and improve solution. In the study, it is found the NN and GA are effective to solve the ODP problems. Both the heuristics mechanism and PEM are useful for improving the performance of NN and GA. GA is more effective and efficient than NN, particularly when the problem scale becomes large. With respect to implementation, the optimization problems in DPMS can be solved in terms of intelligent services, which may be developed using the proposed HMVM model on the GRAM platform.
Subjects: Hong Kong Polytechnic University -- Dissertations
Production engineering
Product management
Computer integrated manufacturing systems
Pages: xiv, 168, [26] leaves : ill. ; 30 cm
Appears in Collections:Thesis

Show full item record

Page views

42
Last Week
1
Last month
Citations as of Mar 24, 2024

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.