09.09.2014 Views

13th International Conference on Membrane Computing - MTA Sztaki

13th International Conference on Membrane Computing - MTA Sztaki

13th International Conference on Membrane Computing - MTA Sztaki

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

R. Freund, I. Pérez-Hurtado, A. Riscos-Núñez, S. Verlan<br />

The first two cases corresp<strong>on</strong>d to those ids j for which the corresp<strong>on</strong>ding<br />

cells should be deleted. We remark that for any p k such that p k ≠ i k , there<br />

is a value z ∈ N ∪ {∗} such that there is a sequence x 1 , . . . , x m with x 1 = p k ,<br />

x m = z, and x j = p xj−1 , 2 ≤ j ≤ m. We denote this by z = last(x). The<br />

above affirmati<strong>on</strong> follows from the fact that the Delete–and–Move relati<strong>on</strong><br />

(c<strong>on</strong>sidered as a parent relati<strong>on</strong>) induces a forest <strong>on</strong> the ids of the cells that<br />

should be deleted. The roots of the obtained trees are given by the functi<strong>on</strong><br />

last and they will collect the objects from all the cells in the tree (if they<br />

are different from ∗).<br />

Next we describe how the c<strong>on</strong>tents is moved:<br />

L 4 = {(i 1 , l 1, ′ w 1 ′′ ) . . . (i n , l n, ′ w n)} ′′ where (i k , l<br />

k ′ , w′ k ) ∈ L 3, 1 ≤ k ≤ n, and<br />

w ′′<br />

j = w ′ j +<br />

⋃<br />

last(k)=j<br />

The deleti<strong>on</strong> of cells induces changes to the relati<strong>on</strong> ρ. We collect these<br />

modificati<strong>on</strong>s as a graph transducer DELET E–NODES that will be run<br />

after the Change–Relati<strong>on</strong> transducer. This transducer deletes all vertices<br />

j such that p j ≠ j as well as all edges that are incoming to these deleted<br />

nodes.<br />

We also remove the corresp<strong>on</strong>ding cells from L 4 :<br />

L 5 = (i 1 , l 1, ′ w 1 ′′ ) . . . (i n1 , l n ′ 1<br />

, w n ′′<br />

1<br />

) where (i j , l j ′ , w′′ j ) ∈ L 4 and p j = i j .<br />

6. (relati<strong>on</strong> change) The new relati<strong>on</strong> C ρ ′ is computed by running the graph<br />

transducers CREAT E–NODES, Change–Relati<strong>on</strong>(r〈v k 〉) and<br />

DELET E–NODES for all (r k , v k ) ∈ R <strong>on</strong> C ρ .<br />

The output of the algorithm is the c<strong>on</strong>figurati<strong>on</strong> Apply(RI, C) = (L 5 , C ′ ρ).<br />

w ′ k.<br />

3 Tax<strong>on</strong>omy<br />

In order to simplify the notati<strong>on</strong>, instead of using a l<strong>on</strong>g tuple bringing everything<br />

together we shall c<strong>on</strong>sider several variants of rule notati<strong>on</strong>, adapted for<br />

the following specific types of rules:<br />

Simple rewriting rule (R-rule)<br />

An R-rule is defined <strong>on</strong>ly by the following comp<strong>on</strong>ents:<br />

r = (Labels(r), ρ(r), Rewrite(r))<br />

Simple rewriting rule with label rename (LR-rule)<br />

An LR-rule is defined <strong>on</strong>ly by the following comp<strong>on</strong>ents:<br />

r = (Labels(r), ρ(r), Rewrite(r), Label–Rename(r))<br />

Simple creati<strong>on</strong> rule (C-rule)<br />

A C-rule is defined by the following comp<strong>on</strong>ents:<br />

r = (Labels(r), ρ(r), Generate(r), Generate–and–Copy(r),<br />

Change–Relati<strong>on</strong>(r))<br />

206

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!