资源分配系统禁止状态的避免研究

On the forbidden state avoidance of resource allocation systems

基于对有界、被P不变式覆盖且P不变式被初始标记的Petri网的结构分析,提出一种避免资源分配系统禁止状态的方法.该方法首先离线计算出资源分配系统Petri网模型中的死标识、坏标识和危险标识等特殊标识,进而在线保证系统不到达死标识和坏标识对应的禁止状态.该方法与其他方法相比,它使系统的许可行为最大化且无需计算网系统的整个可达图,更适合比较庞大的资源分配系统.

Developed in this paper is a new strategy for avoiding forbidden states of resource allocation systems. The strategy is based on Petri nets and their structural analysis, with the Petri nets bounded and covered by initial marked P-invariants. This approach is first to orderly conduct off-line computations of some special states or markings, such as dead （lock） markings, bad markings, and dangerous markings, of the system＇s Petri net model, and then make on-line the system never reach a forbidden state such as dead（lock） markings and bad markings. The major advantage of the technique lies in the fact taht it is suitable for much larger resource allocation systems than the most existing ones and the supervisory control is maximally permissive. Meanwhile, the computation is successfully avoided for the whole teachability graph of the Petri net of the system.