| Author: | Song, Bilian |
| Title: | Intelligent real time optimization system for line-balancing control in apparel manufacture |
| Degree: | Ph.D. |
| Year: | 2008 |
| Subject: | Hong Kong Polytechnic University -- Dissertations. Clothing trade. Production management -- Data processing. |
| Department: | Institute of Textiles and Clothing |
| Pages: | xxx, 254 leaves : ill. ; 30 cm. |
| Language: | English |
| Abstract: | The assembly line configuration and balancing control in apparel manufacture relies heavily on human judgment, although with the advanced systems installed. Consistent decisions and optimal solutions are difficult to obtain and/or maintain under dynamic and uncertain manufacturing environments. The decision-making process is further complicated by the human factors of operatives. An Intelligent Real-time Optimization Decision Support System (IRODSS) is thus developed to assist the supervisors for assembly line balancing control by providing optimal resources re-allocation solutions considering the impact of operator efficiency variance and other dynamic factors. A general assembly line balancing (GALB) problem is addressed in this thesis for an automatic unit production system (AUPS) which produces single products in stochastic processing time with a hybrid line structure. The optimization aims to obtain the following objectives: "to maximize the line efficiency, to minimize the standard deviation of operation efficiency and to minimize the operation efficiency waste". Two stages of optimization, namely, pre-line balancing prior to production and real time line balancing during production, are proposed for fulfilling the above objectives. Pre-line balancing control is initialized by an optimal configuration through an in-depth study of the systems factors on the production process, while real-time balancing control is handled by sequential optimization based on the prediction of real-time operator efficiency. Nowadays the assembly systems in apparel companies are associated with considerable investment cost due to the high level of automation. The optimal configuration of an assembly line is of critical importance for implementing a cost efficient production system. A study on assembly line configuration optimization is thus carried out by analyzing impacts of relevant factors on production flow and line efficiency and reasoning rules for parameter configuration, workstation layout and operator allocation using validated simulation models. Accurate forecasting of operator efficiency is crucial for supervisors in making operator allocation decisions scientifically. A novel time series based artificial neural network (TSANN) model is thus proposed for the purpose of operator efficiency prediction. The TSANN model combines the merits of traditional time series forecasting methods with those of artificial neural network. It is proved to outperform four traditional time series methods in forecasting data patterns of operator efficiency presented in this study. Optimal operator allocation may lower the variance of operation efficiency and hence improve the balancing status of a production line. A multi-objective based recursive operator allocation optimization (ROAO) model is thus developed based on the operator efficiency predicted by TSANN. The operator allocation belongs to the NP-hard combinatorial optimization field. The recursive algorithms attempt to eliminate the weakness of the conventional heuristic procedures in modeling the actual conditions of assembly lines with large problem size and high flexibility. The generated optimal operator allocation is used for system initialization first. But even initialized to the optimal balanced status, the assembly line will turn into unbalanced over time due to uncertain operator efficiency. Whenever the imbalance status is beyond the acceptable level, a decision is required for rebalancing the production line. A sequential decision making (SDM) model, thus, is brought forward for solving the continuous line reconfiguration and/or rebalancing problem in real time basis production. All the proposed models are integrated into an Intelligent Real-time Optimization Decision Support System (IRODSS) through interactive cooperation between Promodel software and Delphi applications. A successful way is provided to simulate the difficult logics like sequential decision making and operator efficiency changing during real time basis production. The developed IRODSS can represent the complexity of real situation because it considers different types of uncertainties (e.g. the variance of operator efficiency) that may exist in a dynamic environment. It is an autonomous temporal intelligent system as it is capable of monitoring the change of work environment (i.e. the change of flow line status) and providing time-based decisions to improve the line balance status sequentially and automatically. These features make it different from other expert systems. |
| Rights: | All rights reserved |
| Access: | open access |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| b22392269.pdf | For All Users | 6.33 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/3558