Measure of Cellular Automata
Juan Carlos Lopez jlopez1967@[Link]
The information contained is confidential and is owned by Juan Carlos Lopez
Motivation
This project was created to conduct an experiment to analyze rule 94 to determine why this rule under certain initial conditions produces two types of behaviors: Class 3 and Class 2. Rule 94 = rule 12 or rule 90
Abstract
We present a set of measures, increasing, decreasing, stable, chaotic, which can be calculated from the transition rule, seen as a Boolean function. These measures can be calculated in two ways. first analyzing each term of the truth table and generate compiled by each of the measures that we will call static measures and the other is to run the automata and take statistics of the occurrence of each measure is the dynamic call.
3
Rule 12
This is a stable rule in which after a few iterations you reach a fixed point of class 2.
Rule 90
It's a chaotic rule characterized by typical chaotic patterns: Class 3 .
Rule 94
This is a rule with a pattern that has initial chaotic behavior (Class 3) and later it develops a stable behavior (class 2). Why?
Why?
Upon decomposing Rule 94 using the simplification of boolean functions, one can observe that: | a | b | c |, the neighbor on the left side is. (a) ; (c) is the neighbor on the right side is (c) and the central cell is (b). Rule 94 can be re-written:
o
(a xor c) or not a and b Doing algebraic manipulation leads to o Rule 94 = (a and b) or (a xor
o o
c)
Using minimization of boolean functions
to perform the same procedure with rules 12 and rule 90 we got that
rule 12 = a and b rule 90 = a xor c
When replaced, we obtain that Rule 94 = rule 12 or rule 90
Observation
Noting Rule 94 in the graph we observe that after the first iteration two points are isolated from each other. That while dynamics do not touch, the behavior of rule 94 is completely determined by Rule 90 and Rule 12 seems not to exist, After the iteration 5 the two meet and there commences an interesting phenomenon, i.e. a collaboration of Rules 12 and 90.
Collaboration is born
This begins to act in Rule 12 allowing fixed patterns of 1 central and counteracting the action in this area of rule 90, the latter which continues to operate outside this area generating more chaotic patterns whilst Rule 12 generates a stable pattern, thereby leading to a synergy of chaotic and stable pattern construction:- chaotic (Class 3) and stable (Class 2).
Measure
After this analysis I remained concerned about how to classify the cellular automata based on the decomposition of the transition function using logic minimization techniques.
10
Measures of Cellular Automata
This job search is based on measures to classify the cellular automata The Cellular automata have two sets of behavior measures: a static, based on the analysis of the transition rule and states:
o o o o
And another dynamic that involves the network and execution dynamics, and states: - Complexity (i.e the resulting action is totally different to the
initial thought) - Similarity 11
Stability Growth Decrease Chaotic
What I see
1. I need to express the rule as a Minimal Boolean Expression 1. I need to determine the Measures that express:
Chaotic behavior Stable behavior Growth Decrease
12
Basic behaviors
4 types of basic behaviors were observed regarding the transition rules. These are as indicated.
Stability B Rule 204
AND
Decrease
A and C Rule 160
OR
Growth
A or B Rule 252
XOR
Growth and Chaotic
A xor C Rule 90
14
Logic minimization
Logic minimization is used to extract features for each term added to the execution of the dynamics Usually when expressing the logic with operators AND, NOT, OR or AND, NOT, XOR, the way to really simplify the expression of this rule is by using AND, NOT and OR. However, I also use XOR which is not meant to eliminate redundancy but rather to rewrite the rule giving it the expression that I need for this. Eg: this rule NOT X AND Y OR X NOT Y CAN BE EXPRESSED BY X XOR Y Because XOR includes within the rule simplified encoding o Note: That chaotic rules are characterized by their expression within the XOR.. I am using the XOR to isolate the chaotic patterns. o There is a pattern behind the XOR that generates triangles behavior patterns that characterize the chaotic and the work was to isolate this pattern to detect if it is in the presence of a chaotic pattern.
15
Which is the scheme used to determine which type of Measures are used?
Operator A Operator B
Operator A 0 0 1 Operator B 0 1 0 1 0 1 0 1 stability decrease decrease unstable unstable growth growth stability Result 0 0 0 0 1 1 1 1
Operatio n Result
Example a 0 0 1 1 c 0 1 0 1 a xor c 0 stability 1 growth 1 decrease 0 chaotic
1 0 0 1 1
Because operations can be applied in pairs to achieve an indicator that can apply 16any rule, giving the to same result
Structure of rule 94
111 110 101 100 011 010 001 000 0 0 0 1 1 1 1 0
(q AND (NOT p)) OR (p XOR r)
17
Evaluation each term, Begining with 111
111 0 110 0 101 0 100 1 011 1 010 1 001 1 000 0
Result stability 0
Operator A Operator B 0 0 1 1 0 0 1 1 0 1 0 1 0 1 0 1
decrease 0 decrease 0 unstable 0
unstable 1 growth growth 1 1
stability 1
111 0(Chaotic)
110 0
101 0
100 1
011 1
010 1
18
001 1
000 0
Evaluation each term, Begining with 111
111 0 110 0 101 0 100 1 011 1 010 1 001 1 000 0
Result Operator A Operator B 0 0 1 1 0 0 1 1
The way tostability 0 move 0 actions between nodes 1 decrease 0 is a heuristics, 0 decrease 0 developed by myself
1 0 1 0 1 unstable 0
unstable 1 growth growth 1 1
stability 1
111
110
101
100
011
010
001 19
000
Show the table and measure by the rule 90
111 0(Chaotic) 110 1(Grow) 101 0(Chaotic) 100 1(Grow) 011 1(Grow) 010 0(Stable) 001 1(Grow) 000 0(Stable)
Count the patters and compute the occurrence of each measure
20
Take static measurements of Cellular Automata
In Rule 90 the following measures can be observed: 25% stability, 50% growth 25% chaotic .
This is how I determine the static measures:
I put these factors into a vector so we have static (stability, growth, decrease, chaotic) We still have that rule 90 is matched to Rule 90 static(25%, 50%, 0%, 25%)
21
What are the Measures that determine this chaotic behavior
After many experiments I was able to isolate and determine which is the pattern that generates the chaotic pattern: the red cell identifies the chaotic pattern; the black cell identifies the growing pattern. Note that afterwards, you receive a pattern that is formed where the triangles characterize chaos.
Rule 90
22
Dynamic Measurement
So far there has been cellular automata transition rule but there are other characteristics involved in the dynamics of the C.A., which are the grid and the initial state. Cellular Automata have their nature which we observed in the previous section on static measures however, the property of the grid and the initial state, determine the complexity of cellular automata after running the automaton dynamics By putting now the dynamic statistics in a vector by first: - running the dynamics, then - compiling the statistics, and finally - grouping them into a vector
o
Dynamic (stability, growth, decline, chaotic) accounting for the patterns.
(An example is on the following page.)
23
With this table execute the dynamic and take statical
111 0(Chaotic) 110 1(Grow) 101 0(Chaotic) 100 1(Grow) 011 1(Grow) 010 0(Stable) 001 1(Grow) 000 0(Stable)
Count the cell and compute the occurrence of each measure, and generate the % global, it generate the dynamic measure
24
With this table execute the dynamic and take statical
111 0(Chaotic) 110 1(Growth) 101 0(Chaotic) 100 1(Growth) 011 1(Growth) 010 0(Stable) 001 1(Growth) 000 0(Stable)
Count the cell and compute the occurrence of each measure, and generate the % global, it generate the dynamic measure
I put these factors into a vector so we have dynamic (stability, growth, decrease, chaotic) We still have that rule 90 is matched to Rule 90 dynamic(24.2%,50.8 %, 0%, 25%) 25
Measuring the complexity
The experiments show that the static and dynamic measurements are very similar in most cases, however there are a set of CA as observed in the game of life that when analyzing the static and dynamic measurements, their comportment is very different or wide apart. Complexity is inversely proportionate to the correlation between static and dynamic measures.
27
Analysis the game of life
Measure Static Decrease Chaotic Growth Stability 4,68 67,96 27,34 0,00 Measure Dinamic 75,23 13,38 11,37 0,00 Differe nce 70,55 -54,58 -15,97 0
Correlation
-0,29
28
Analysis the game of life
Measure Static Decrease Chaotic Growth Stability 4,68 67,96 27,34 0,00 Measure Dinamic 75,23 13,38 11,37 0,00 Differe nce 70,55 -54,58 -15,97 0
Correlation
-0,29
29
Similarities between the game of life and Rule 94
These patterns are similar both are in a stable pattern Environment where it coexists With Instability
30
My Hypothesis
My hypothesis is that Cellular Automata like the game of life, have a complex behavior because they have much potential chaos that is being repressed by stable patterns similar to the behavior discussed in the beginning where chaos rules 94 whereas the central pattern was controlled by stability. Such as a stable structure contains a potential energy that when colliding with anything repressed a blast is produced making these structures change shape and thereby generating a totally different dynamic to that prior to the collision.
31
Characteristics of metrics
It depends on the crosslinked You can apply to all binary cellular automata When searching complex rules similar to the game of life, I select those that are most likely, by analyzing the transition rule One can learn also from the analysis of the rule which will be the expected behavior for a cellular automaton, cellular automata such as class 1 or class 2 are those that do not have the chaotic factor
32
Conclusions
By its nature the AC can be analyzed qualitatively from the transition rule to observe 4 metric (stability, growth, decrease, chaotic). The complexity stems from the environment where you run the cellular automata such as the initial state of CA and the grid on which to run this automatically. These factors directly affect the complexity of the cellular automata; to measure the impact of these factors one should run dynamics, take the statistics of occurrence of different factors obtainable after several experiments with different initial statesthat can determine how different are the measures of the dynamics static. For example if we take the rule of the AC game of life and made a small change to its structure as easy as safe execution of the automata it is totally different to the original AC becoming a chaotic as the static characteristics. The papers give metric patterns with which the AC can be classified qualitatively based on the simplification of regulation of AC.
34
Rule of Life game after of processing minimization logic
~x0&~x1&([x2]^[x3])&([x4]^[x5])&([x6]^[x7])| ~x0&~x2&([x1]^[x3])&([x4]^[x5])&([x6]^[x7])| ~x1&~x2&([x0]^[x3])&([x4]^[x5])&([x6]^[x7])| ~x0&~x1&([x2]^[x4])&([x3]^[x5])&([x6]^[x7])| ~x0&~x2&([x1]^[x4])&([x3]^[x5])&([x6]^[x7])| ~x1&~x2&([x0]^[x4])&([x3]^[x5])&([x6]^[x7])| ~x0&~x3&([x1]^[x4])&([x2]^[x5])&([x6]^[x7])| ~x1&~x3&([x0]^[x4])&([x2]^[x5])&([x6]^[x7])| ~x2&~x3&([x0]^[x4])&([x1]^[x5])&([x6]^[x7])| ~x0&~x1&([x2]^[x3])&([x4]^[x6])&([x5]^[x7])| ~x0&~x2&([x1]^[x3])&([x4]^[x6])&([x5]^[x7])| ~x1&~x2&([x0]^[x3])&([x4]^[x6])&([x5]^[x7])| ~x0&~x1&([x2]^[x4])&([x3]^[x6])&([x5]^[x7])| ~x0&~x2&([x1]^[x4])&([x3]^[x6])&([x5]^[x7])| ~x1&~x2&([x0]^[x4])&([x3]^[x6])&([x5]^[x7])| ~x0&~x3&([x1]^[x4])&([x2]^[x6])&([x5]^[x7])| ~x1&~x3&([x0]^[x4])&([x2]^[x6])&([x5]^[x7])| ~x2&~x3&([x0]^[x4])&([x1]^[x6])&([x5]^[x7])| ~x0&~x1&([x2]^[x5])&([x3]^[x6])&([x4]^[x7])| ~x0&~x2&([x1]^[x5])&([x3]^[x6])&([x4]^[x7])| ~x1&~x2&([x0]^[x5])&([x3]^[x6])&([x4]^[x7])| ~x0&~x3&([x1]^[x5])&([x2]^[x6])&([x4]^[x7])| ~x1&~x3&([x0]^[x5])&([x2]^[x6])&([x4]^[x7])| ~x2&~x3&([x0]^[x5])&([x1]^[x6])&([x4]^[x7])| ~x0&~x4&([x1]^[x5])&([x2]^[x6])&([x3]^[x7])| ~x1&~x4&([x0]^[x5])&([x2]^[x6])&([x3]^[x7])| ~x2&~x4&([x0]^[x5])&([x1]^[x6])&([x3]^[x7])| ~x3&~x4&([x0]^[x5])&([x1]^[x6])&([x2]^[x7])| ~x0&~x1&~x2&([x3]^[x4])&([x5]^[x6])&x8| ~x0&~x1&~x3&([x2]^[x4])&([x5]^[x6])&x8| ~x0&~x2&~x3&([x1]^[x4])&([x5]^[x6])&x8| ~x1&~x2&~x3&([x0]^[x4])&([x5]^[x6])&x8| ~x0&~x1&~x2&([x3]^[x5])&([x4]^[x6])&x8| ~x0&~x1&~x3&([x2]^[x5])&([x4]^[x6])&x8| ~x0&~x2&~x3&([x1]^[x5])&([x4]^[x6])&x8| ~x1&~x2&~x3&([x0]^[x5])&([x4]^[x6])&x8| ~x0&~x1&~x4&([x2]^[x5])&([x3]^[x6])&x8| ~x0&~x2&~x4&([x1]^[x5])&([x3]^[x6])&x8| ~x1&~x2&~x4&([x0]^[x5])&([x3]^[x6])&x8| ~x0&~x3&~x4&([x1]^[x5])&([x2]^[x6])&x8| ~x1&~x3&~x4&([x0]^[x5])&([x2]^[x6])&x8| ~x2&~x3&~x4&([x0]^[x5])&([x1]^[x6])&x8| ~x0&~x1&~x2&([x3]^[x4])&([x5]^[x7])&x8| ~x0&~x1&~x3&([x2]^[x4])&([x5]^[x7])&x8| ~x0&~x2&~x3&([x1]^[x4])&([x5]^[x7])&x8| ~x1&~x2&~x3&([x0]^[x4])&([x5]^[x7])&x8| ~x0&~x1&~x2&([x3]^[x5])&([x4]^[x7])&x8| ~x0&~x1&~x2&([x3]^[x6])&([x4]^[x7])&x8|
35
