Abstract
The original protection graph rewriting grammar used to simulate the different operations of the Take/Grant model is reviewed. The productions of the PGR-grammar is then expanded, by adding a new context which is based on the different security classes found in the Bell and LaPadula model.