科研之友
微信
新浪微博
Facebook
分享链接
摘要
A simulation-based approach is one of the most effective ways of solving design problems of production systems under uncertain conditions. However, a simulator generating system is indispensable to employ this approach. Also, such a generator is required to devise an appropriate operational rule set for each simulator. In this work, a design support system was developed using object-oriented technique, which composes simulators with various configurations of software components. Reinforcement learning was adopted in order to obtain a rule set for operating each of composed simulators. Verification and evaluation of effectiveness of the method was confirmed through a series of experiments (design of lanes and locations of machines, acquisition of proper control rules of vehicles) using an automated guarded vehicle system (AGVS) as a model system.
- 出版日期2010
