Multi-objective Optimization of AGV Centralized Control System

• Main Authors: Chan, Ching-Hsiang, 詹清翔
• Other Authors: 王小璠
• Format: Others
• Language: en_US
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/14306404409627121827

碩士 === 國立清華大學 === 工業工程與工程管理學系 === 102 === Due to the rise of labor cost and the requirement of quality stability, industrial automation has become a solution for enterprises to strengthen the competitiveness and reduce the operational costs in recent years. In the field of industrial automation, automated guided vehicle (AGV) system has been generally adopted for transporting semi-manufactured goods, components or products between workstations to facilitate production processes in an unmanned factory. In order to satisfy production processes, the planned routes of AGVs which coordinate workstations and a warehouse play a critical rule to enhance system performance. This study first investigates the important factors which affect the vehicle routing performances of the AGVs and has the following observations: (1) In-time material delivery along workstations smoothens the production process; (2) The lighter “en route vehicle weight” alleviates energy wastes; (3) The increased space utilization of an AGV reduces the fleet size of AGVs. Based on the aforementioned analysis, a mathematical model in the form of multi-objective integer programming model is developed to obtain the optimal vehicle routings for pick-ups and deliveries of multiple products/components along the production flows such that energy consumption and operational cost can be minimized. Moreover, in order to have further applications in a practical circumstance, the simultaneously dispatching and routing algorithm, called Intelligent Control Algorithm, is proposed to enhance system performance, such as the moving distance and the completion time of the products. Finally, an illustrative example from an assembly company of Taiwan is presented to verify our proposed deterministic model and dynamic algorithm embedded in the simulation software, Flexsim. With given layout and production flows, the result shows that the pick-up and delivery routes of AGVs can be determined optimally and ensured feasibility with sufficient fleet in deterministic model and our proposed rules based on the algorithm can outperform the conventional nearest-vehicle dispatching rules.