Editing P system (section)

===Rule application===
At each step of a computation an object may only be used once, as they are consumed by rules when applied. The method of applying a rule within a membrane is as follows: 
#Assign symbols from a membrane's content to the rule's inputs
#If all inputs are satisfied, remove all assigned symbols from membrane 
#Create output symbols and hold until all rule assignment, for all membranes, has taken place. 
#Add output symbols to targeted membranes. 
#Dissolve membranes as necessary 
Outputs are not passed immediately into membranes because this would contravene the maximally parallel nature of rule application, instead they are distributed after all possible rules have been applied.

====Non-deterministic application====
The order of rule application is chosen at random. Rule application order can have a significant effect on which rules may be applied at any given time, and the outcome of a step of execution.

Consider a membrane containing only a single "a" symbol, and the two rules a → ab and a → aδ. As both rules rely on an “a” symbol being present, of which there is only one, the first step of computation will allow either the first or second rule to be applied, but not both. The two possible results of this step are very different:
#The membrane carries over to the next step of the computation with both an "a" symbol and a "b" symbol present, and again one of the two rules is randomly assigned to the "a" symbol.
#The membrane dissolves and a single "a" symbol is passed out to the containing membrane.

====Maximally parallel application====
This is a property of rule application whereby all possible rule assignments must take place during every step of the computation. In essence this means that the rule a → aa has the effect of doubling the number of "a" symbols in its containing membrane each step, because the rule is applied to every occurrence of an "a" symbol present.