Deadlock prevention using divide-and-conquer strategy for WS3PR

摘要

This work applies a divide-and-conquer strategy to handle deadlock prevention issues in a class of generalized Petri nets, WS3PR, an extended S3PR. And, the strategy is improved by utilizing a previous result of WS3PR, i.e., a WS3PR with a characteristic structure is inherently live at some proper initial markings. Given a marked WS3PR, after being divided into an idle subnet, an autonomous one, and some so-called toparchies, every toparchy is checked according to the result to decide whether it is inherently live at the current initial marking. This strategy for deadlock prevention in WS3PR can effectively reduce computational complexity, structural complexity, and improve system permissiveness. Some explanatory examples are used to illustrate this strategy.