基于Petri网下哲学家就餐问题模型的改进与推广  

Improvement and Extension of the Dining Philosophers Problem Based on Petri Nets

作　　者：邵长明[1] 宋文[1] 

机构地区：[1]西华大学数学与计算机学院,成都610039

出　　处：《计算机与数字工程》2010年第1期40-43,共4页Computer & Digital Engineering

摘　　要：哲学家就餐问题是描述在共享资源下同步与并发的经典案例,活性与无饥饿性是求解此问题的前提,效率是基本要求。基于两个许可卡的传递策略,已经给出了无饥饿解的Petri网模型,并且在该模型基础上的改进解决了许可卡的传递问题同时也解决了拥有许可卡且准备就餐的两位相邻哲学家之间共享资源的冲突问题。然而以上改进并没有真正实现效率上的高效且不易推广到哲学家人数大于5位的情形。The dining philosophers problem is a classical example of synchronization and concurrency of describing the shared resources. The live and starvation--free are the premise to solve this problem, and efficiency is basic requirements. A Petri net--based starvation--free solution was described, in which two dining--cards were cycled, and the improvment based on this model solves the problem of two cards＇ cycled and the conflict of two adjacent philosophers in shareing resource who has the card and are ready for dining. However, these improvements do not truly improve the efficiency, and it is impossible to be extended to the cases of number which is greater than 5.

关 键 词：PETRI网 哲学家就餐问题 延迟Petri网 效率 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]