A Characterisation of NL Using Membrane Systems without Charges and Dissolution
25 Aug 2008-pp 164-176
TL;DR: It turns out that the computational power of some systems is lowered from P to NL, so it seems that the tighter uniformities are more reasonable for these systems.
Abstract: We apply techniques from complexity theory to a model of biological cellular membranes known as membrane systems or P-systems. Like circuits, membrane systems are defined as uniform families. To date, polynomial time uniformity has been the accepted uniformity notion for membrane systems. Here, we introduce the idea of using AC 0and L -uniformities and investigate the computational power of membrane systems under these tighter conditions. It turns out that the computational power of some systems is lowered from P to NL , so it seems that our tighter uniformities are more reasonable for these systems. Interestingly, other systems that are known to be lower bounded by P are shown to retain their computational power under the new uniformity conditions. Similarly, a number of membrane systems that are lower bounded by PSPACE retain their power under the new uniformity conditions.
Summary (2 min read)
Jump to: [1 Introduction] – [2 Membrane Systems] – [2.1 Recogniser membrane systems] – [2.2 Complexity classes] – [3 NL Upper bound on active membranes without dissolution rules] – [4 NL lower bound for semi-uniform active membranes without dissolution] – [4.1 PARITY lower bound for uniform active membranes without dissolution] and [5 P lower bound on uniform families of active membrane systems with dissolving rules]
1 Introduction
- Membrane systems [12] are a model of computation inspired by living cells.
- In the active membrane model it is also possible for a membrane to completely dissolve, and for a membrane to divide into two child membranes.
- The authors also show that the PSPACE lower and upper bounds mentioned above still hold under these restricted uniformity conditions.
- So far the authors have shown that four models, that characterise P when polynomial time uniformity is used, actually only characterise NL when restricted to be AC0 uniform.
- Interestingly, the authors also show that two other polynomial time uniform membrane system that are known [9] to be lower bounded by P actually retain this P lower bound when restricted to be AC0 uniform.
2 Membrane Systems
- In this section the authors define membrane systems and complexity classes.
- The authors also introduce the notion of AC0 uniformity for membrane systems.
2.1 Recogniser membrane systems
- Active membranes systems are membrane systems with membrane division rules.
- Division rules can either only act on elementary membranes, or else on both elementary and non-elementary membranes.
- These rules are applied according to the following principles: – All the rules are applied in maximally parallel manner.
- The result of the computation (a solution to the instance) is “yes” if a distinguished object yes is expelled during the computation, otherwise the result is “no”.
- Therefore, the following interpretation holds: given a fixed initial configuration, a confluent membrane system nondeterministically chooses one from a number of valid configuration sequences, but all of them must lead to the same result.
2.2 Complexity classes
- Here the authors introduce the notion of AC0 uniformity to membrane systems.
- Previous work on the computational complexity of membrane systems used (Turing machine) polynomial time uniformity [14].
- – Each ΠX(n) is confluent: all computations of ΠX(n) with the same input x of size n give the same result; either always “yes” or else always “no”.
- The authors denote by AM0+ne the classes of membrane systems with active membranes, and both non-elementary and elementary membrane division and no charges.
- The authors now show that the use of AC0 uniformity does not change this lower bound.
3 NL Upper bound on active membranes without dissolution rules
- Previously the upper bound on all active membrane systems without dissolution was P [5].
- The authors give an overview rather than the full details.
- The authors make the observation that the graph GΠ can be constructed in deterministic logspace, and even in AC0.
- Since the authors have shown that the problem of simulating a membrane system without charges and without dissolution can be encoded as an NL-complete problem they have proved Theorem 1.
4 NL lower bound for semi-uniform active membranes without dissolution
- The algorithm works by having each edge in the problem instance graph represented as a membrane.
- The initial multisets are all empty except Mcount = {c2n+1}.
- The authors also have a counter that counts down in parallel with the above steps.
- Note that the authors encode the edges of the graph as membranes, rather than objects.
- In the membrane computing framework, for uniform membrane systems, inputs must be specified as objects.
4.1 PARITY lower bound for uniform active membranes without dissolution
- The previous proof gave a lower bound for a semi-uniform membrane system.
- The authors show that PARITY ∈ PMCAM0−d,+u by providing an AC0 uniform membrane system that can solve instances of the problem.
- PARITYis the problem of telling whether the number of 1 symbols in the input word is odd.
- A type (a) rule is created mapping every even object with i “1” symbols to the odd object with i− 1 “1” symbols in it.
- The AC0 uniformity machine (a CRAM) rearranges the input word w by moving all 1 symbols to the left and all 0 symbols to the right, to give w′.
5 P lower bound on uniform families of active membrane systems with dissolving rules
- In this section the authors show that does not happen for all models with at least P power.
- Naturally this result also holds for the semi-uniform case.
- The resulting membrane system directly solves the instance of CVP in polynomial time.
- The authors simulate multiple fanouts by outputting multiple copies of the resulting truth value of each gate.
- The output of a gate moves up through the layers of the membrane system until it reaches the correct gate according to its tag.
Did you find this useful? Give us your feedback
A characterisation of NL using membrane
systems without charges and dissolution
Niall Murphy, Damien Woods
Technical Report
NUIM-CS-TR-2008-01
Department of Computer Science
National University of Ireland, Maynooth
Ireland
A characterisation of NL using membrane
systems without charges and dissolution
Niall Murphy
1
and Damien Woods
2
1
Department of Computer Science, National University of Ireland, Maynooth,
Ireland
nmurphy@cs.nuim.ie
2
Department of Computer Science, University College Cork, Ireland
d.woods@cs.ucc.ie
Abstract. We apply techniques from complexity theory to a model of
biological cellular membranes known as membrane systems or P-systems.
Like circuits, membrane systems are defined as uniform families. To
date, polynomial time uniformity was the accepted uniformity notion
for membrane systems. Here, we introduce the idea of using AC
0
and L
uniformities and investigate the computational power of membrane sys-
tems under these tighter conditions. It turns out that the computational
power of some systems is lowered from P to NL, so it seems that our
tighter uniformities are more reasonable for these systems. Interestingly,
other systems that are known to be lower bounded by P are shown to
retain their computational power under the new uniformity conditions.
Similarly, a number of membrane systems that are lower bounded by
PSPACE retain their power under the new uniformity conditions.
1 Introduction
Membrane systems [12] are a model of computation inspired by living cells.
In this paper we explore the computational power of cell division (mitosis) and
dissolution (apoptosis) by investigating a variant of the model called active mem-
branes [11]. An instance of the model consists of a number of (possibly nested)
membranes, or compartments, which themselves contain objects. During a com-
putation, the objects, depending on the compartment they are in, become other
objects or pass through membranes. In the active membrane model it is also
possible for a membrane to completely dissolve, and for a membrane to divide
into two child membranes.
This membrane model can be regarded as a model of parallel computation
but it has a number of features that make it somewhat unusual when compared
to other parallel models. For example, object interactions are nondeterministic
so confluence plays an important role, membranes contain multisets of objects,
there are many parameters to the model, etc. In order to clearly see the power
of the model we analyse it from the computational complexity point of view, the
goal being to characterise the model in terms of the set of problems that it can
solve in reasonable time.
Another, more specific, motivation is the so-called P-conjecture [13] which
states that recogniser membranes systems with division rules (active membranes),
but without charges, characterise P. On the one hand, it was shown that this
conjecture does not hold for systems with non-elementary division as PSPACE
upper [16] and lower [1] bounds were found for this variant (non-elementary di-
vision is where a membrane containing multiple membranes and objects may be
copied in a single timestep). On the other hand, the P-conjecture was shown to
hold for all active membrane systems without dissolution rules, when Guti´errez-
Naranjo et al. [5] gave a P upper bound. The corresponding P lower bound
(trivially) came from the fact that the model is defined to be P-uniform.
However, here we argue that the aforementioned P lower bound highlights a
problem with using P uniformity, as it does not tell us whether this membrane
model itself has (in some sense) the ability to solve all of P in polynomial time,
or if the uniformity condition is providing the power. In this paper we show that
in fact when we use weaker, and more reasonable, uniformity conditions the
model does not have the ability to solve all problems in P (assuming P 6= NL).
We find that with either AC
0
or L uniformity the model characterises NL in
the semi-uniform case, and we give an NL upper bound for the uniform case.
We also show that the PSPACE lower and upper bounds mentioned above still
hold under these restricted uniformity conditions.
Using the notation of membrane systems (to be defined later) our upper
bound on L-uniform and L-semi-uniform membrane systems can be stated as
follows.
Theorem 1. PMC
AM
0
−d
⊆ NL
Essentially this theorem states that polynomial time active membrane systems,
without dissolution rules, solve no more than those problems in NL. Despite
the fact that these systems run for polynomial time (and can even create expo-
nentially many objects), they can not solve all of P (assuming NL 6= P). This
result is illustrated by the bottom four nodes in Figure 1.
The upper bound in Theorem 1 is found by showing that the construction
in [5] can be reduced to an instance of the NL-complete problem s-t-connectivity
(STCON). The full proof appears in Section 3. Next we give a corresponding
lower bound.
Theorem 2. NL ⊆ PMC
AM
0
−d,−u
To show this lower bound we provide an AC
0
-semi-uniform membrane family
that solves STCON. The full proof is in Section 4 and the result is illustrated
by the bottom left two nodes in Figure 1. Therefore, in the semi-uniform case
we have a characterisation of NL.
Corollary 1. NL = PMC
AM
0
−d,−u
We have not yet shown an analogous lower bound result for uniform families. To
date our best lower bound is PARITY, which is known not to be in AC
0
[4].
We describe this in Section 4.1.
NL
PSPACE
P
PSPACE
P
NL
PARITY
NL
NL
PARITY
PSPACE PSPACE
-d, -ne, -u
+d, -ne, -u +d, -ne, +u
-d, -ne, +u
-d, +ne, -u -d, +ne, +u
+d, +ne, -u +d, +ne, +u
Fig. 1. An inclusion diagram showing the currently known upper and lower bounds
on the variations of the model. The top part of a node represents the best known
upper bounds, and the lower part the best known lower bounds. An undivided node
represents a characterisation.
So far we have shown that four models, that characterise P when polynomial
time uniformity is used, actually only characterise NL when restricted to be AC
0
uniform. Interestingly, we also show that two other polynomial time uniform
membrane system that are known [9] to be lower bounded by P actually retain
this P lower bound when restricted to be AC
0
uniform. This result is stated as
a P lower bound on membrane systems with dissolution:
Theorem 3. P ⊆ PMC
AM
0
+d,+u
The proof appears in Section 5 and is illustrated by the top front two nodes in
Figure 1.
In Section 2.3 we observe that the known PSPACE upper and lower bounds
(top four nodes in Figure 1) remain unchanged under AC
0
uniformity conditions.
2 Membrane Systems
In this section we define membrane systems and complexity classes. These def-
initions are from P˘aun [11, 12], and Sos´ık and Rodr´ıguez-Pat´on [16]. We also
introduce the notion of AC
0
uniformity for membrane systems.
2.1 Recogniser membrane systems
Active membranes systems are membrane systems with membrane division rules.
Division rules can either only act on elementary membranes, or else on both
elementary and non-elementary membranes. An elementary membrane is one
which does not contain other membranes (a leaf node, in tree terminology).
Definition 1. An active membrane system without charges is a tuple Π =
(O, H, µ, w
1
, . . . , w
m
, R) where,
1. m > 1 is the initial number of membranes;
2. O is the alphabet of objects;
3. H is the finite set of labels for the membranes;
4. µ is a membrane structure, consisting of m membranes, labelled with ele-
ments of H;
5. w
1
, . . . , w
m
are strings over O, describing the multisets of objects placed in
the m regions of µ.
6. R is a finite set of developmental rules, of the following forms:
(a) [ a → v ]
h
,
for h ∈ H, a ∈ O, v ∈ O
∗
(b) a[
h
]
h
→ [
h
b ]
h
,
for h ∈ H, a, b ∈ O
(c) [
h
a ]
h
→ [
h
]
h
b,
for h ∈ H, a, b ∈ O
(d) [
h
a ]
h
→ b,
for h ∈ H, a, b ∈ O
(e) [
h
a ]
h
→ [
h
b ]
h
[
h
c ]
h
,
for h ∈ H, a, b, c ∈ O.
(f) [
h
o
[
h
1
]
h
1
[
h
2
]
h
2
[
h
3
]
h
3
]
h
0
→ [
h
0
[
h
1
]
h
1
[
h
3
]
h
3
]
h
0
[
h
0
[
h
2
]
h
2
[
h
3
]
h
3
]
h
0
,
for h
0
, h
1
, h
2
, h
3
∈ H.
These rules are applied according to the following principles:
– All the rules are applied in maximally parallel manner. That is, in one step,
one object of a membrane is used by at most one rule (chosen in a non-
deterministic way), but any object which can evolve by one rule of any form,
must evolve.
– If at the same time a membrane labelled with h is divided by a rule of type
(e) or (f) and there are objects in this membrane which evolve by means
of rules of type (a), then we suppose that first the evolution rules of type
(a) are used, and then the division is produced. This process takes only one
step.
– The rules associated with membranes labelled with h are used for membranes
with that label. At one step, a membrane can be the subject of only one rule
of types (b)-(f).
In this paper we study the language recognising variant of membrane systems
that solves decision problems. A distinguished region contains, at the beginning
of the computation, an input — a description of an instance of a problem. The
result of the computation (a solution to the instance) is “yes” if a distinguished
object yes is expelled during the computation, otherwise the result is “no”. Such
a membrane system is called deterministic if for each input a unique sequence of
configurations exists. A membrane system is called confluent if it always halts
and, starting from the same initial configuration, it always gives the same re-
sult, either always “yes” or always “no”. Therefore, the following interpretation
Citations
More filters
••
20 Aug 2014TL;DR: This paper presents three designs of uniform families of P systems that solve the decision problem STCON by using Membrane Computing strategies (pure Membranes Computing techniques): P systems with membrane creation, P systemswith active membranes with dissolution and without polarizations and P systemsWith active membranes without dissolution and with polarizations.
Abstract: In Membrane Computing, the solution of a decision problem \(X\) belonging to the complexity class P via a polynomially uniform family of recognizer P systems is trivial, since the polynomial encoding of the input can involve the solution of the problem. The design of such solution has one membrane, two objects, two rules and one computation step. Stricto sensu, it is a solution in the framework of Membrane Computing, but it does not use Membrane Computing strategies. In this paper, we present three designs of uniform families of P systems that solve the decision problem STCON by using Membrane Computing strategies (pure Membrane Computing techniques): P systems with membrane creation, P systems with active membranes with dissolution and without polarizations and P systems with active membranes without dissolution and with polarizations. Since STCON is NL-complete, such designs are constructive proofs of the inclusion of NL in \(\mathbf{PMC}_\mathcal{MC}\), \(\mathbf{PMC}_{\mathcal{AM}^0_{+d}}\) and \(\mathbf{PMC}_{\mathcal{AM}^+_{-d}}\).
3 citations
Cites background from "A Characterisation of NL Using Memb..."
...In [6], Murphy and Woods claim that NL ⊆ PMCAM−d,−u , i....
[...]
...According to [6], the P system of the cited model that solves this instance has Γ = {s, t, yes, no, c0, ....
[...]
••
17 Jun 2009TL;DR: In this article, the authors investigate the affect of various acceptance conditions on membrane systems without dissolution and demonstrate that two particular acceptance conditions (one easier to program, the other easier to prove correctness) both characterise the same complexity class, NL.
Abstract: In this paper we investigate the affect of various acceptance conditions on recogniser membrane systems without dissolution. We demonstrate that two particular acceptance conditions (one easier to program, the other easier to prove correctness) both characterise the same complexity class, NL. We also find that by restricting the acceptance conditions we obtain a characterisation of L. We obtain these results by investigating the connectivity properties of dependency graphs that model membrane system computations.
3 citations
01 Jan 2014
TL;DR: In this article, the authors present three designs of uniform families of P systems that solve the decision problem STCON by using pure Membrane Computing strategies: P systems with membrane creation and active membranes with dissolution and without polarizations.
Abstract: In Membrane Computing, the solution of a decision problem \(X\) belonging to the complexity class P via a polynomially uniform family of recognizer P systems is trivial, since the polynomial encoding of the input can involve the solution of the problem. The design of such solution has one membrane, two objects, two rules and one computation step. Stricto sensu, it is a solution in the framework of Membrane Computing, but it does not use Membrane Computing strategies. In this paper, we present three designs of uniform families of P systems that solve the decision problem STCON by using Membrane Computing strategies (pure Membrane Computing techniques): P systems with membrane creation, P systems with active membranes with dissolution and without polarizations and P systems with active membranes without dissolution and with polarizations. Since STCON is NL-complete, such designs are constructive proofs of the inclusion of NL in \(\mathbf{PMC}_\mathcal{MC}\), \(\mathbf{PMC}_{\mathcal{AM}^0_{+d}}\) and \(\mathbf{PMC}_{\mathcal{AM}^+_{-d}}\).
3 citations
01 Jan 2008
TL;DR: In this article, the authors investigate the affect of various acceptance conditions on membrane systems without dissolution and demonstrate that two particular acceptance conditions (one easier to program, the other easier to prove correctness) both characterise the same complexity class, NL.
Abstract: In this paper we investigate the affect of various acceptance conditions on recogniser membrane systems without dissolution. We demonstrate that two particular acceptance conditions (one easier to program, the other easier to prove correctness) both characterise the same complexity class, NL. We also find that by restricting the acceptance conditions we obtain a characterisation of L. We obtain these results by investigating the connectivity properties of dependency graphs that model membrane system computations.
3 citations
01 Jan 2014
TL;DR: This paper explores P systems where the target of the computation is the construction of a new membrane structure according to its set of rules, and focuses on the self-construction of recognizer P systems.
Abstract: Usually, the changes produced in the membrane structure of a P system are considered side effects. The output of the computation is encoded as a multiset placed in a specific region and the membrane structure in the halting configuration is not considered important. In this paper we explore P systems where the target of the computation is the construction of a new membrane structure according its set of rules. The new membrane structure can be considered as the initial one of a new self-constructed P system. We focus on the self-construction of recognizer P systems and illustrates the definition with a study of the self-construction P systems working as decision trees for solving Machine Learning decision problems.
2 citations
References
More filters
•
19 Dec 1990
TL;DR: The Handbook of Theoretical Computer Science provides professionals and students with a comprehensive overview of the main results and developments in this rapidly evolving field.
Abstract: "Of all the books I have covered in the Forum to date, this set is the most unique and possibly the most useful to the SIGACT community, in support both of teaching and research.... The books can be used by anyone wanting simply to gain an understanding of one of these areas, or by someone desiring to be in research in a topic, or by instructors wishing to find timely information on a subject they are teaching outside their major areas of expertise." -- Rocky Ross, "SIGACT News" "This is a reference which has a place in every computer science library." -- Raymond Lauzzana, "Languages of Design" The Handbook of Theoretical Computer Science provides professionals and students with a comprehensive overview of the main results and developments in this rapidly evolving field. Volume A covers models of computation, complexity theory, data structures, and efficient computation in many recognized subdisciplines of theoretical computer science. Volume B takes up the theory of automata and rewriting systems, the foundations of modern programming languages, and logics for program specification and verification, and presents several studies on the theoretic modeling of advanced information processing. The two volumes contain thirty-seven chapters, with extensive chapter references and individual tables of contents for each chapter. There are 5,387 entry subject indexes that include notational symbols, and a list of contributors and affiliations in each volume.
3,089 citations
•
01 Jan 1997
TL;DR: Throughout the book, Sipser builds students' knowledge of conceptual tools used in computer science, the aesthetic sense they need to create elegant systems, and the ability to think through problems on their own.
Abstract: From the Publisher:
Michael Sipser's philosophy in writing this book is simple: make the subject interesting and relevant, and the students will learn. His emphasis on unifying computer science theory - rather than offering a collection of low-level details - sets the book apart, as do his intuitive explanations. Throughout the book, Sipser - a noted authority on the theory of computation - builds students' knowledge of conceptual tools used in computer science, the aesthetic sense they need to create elegant systems, and the ability to think through problems on their own. INTRODUCTION TO THE THEORY OF COMPUTATION provides a mathematical treatment of computation theory grounded in theorems and proofs. Proofs are presented with a "proof idea" component to reveal the concepts underpinning the formalism. Algorithms are presented using prose instead of pseudocode to focus attention on the algorithms themselves, rather than on specific computational models. Topic coverage, terminology, and order of presentation are traditional for an upper-level course in computer science theory. Users of the Preliminary Edition (now out of print) will be interested to note several new chapters on complexity theory: Chapter 8 on space complexity; Chapter 9 on provable intractability, and Chapter 10 on advanced topics, including approximation algorithms, alternation, interactive proof systems, cryptography, and parallel computing.
2,877 citations
"A Characterisation of NL Using Memb..." refers methods in this paper
...STCON is also known as PATH [17] and REACHABILITY [12]....
[...]
•
01 Jan 2002
TL;DR: This chapter discusses Membrane Computing, What It Is and What It is Not, and attempts to get back to reality with open problems and Universality results.
Abstract: Preface.- 1. Introduction: Membrane Computing, What It Is and What It Is Not.- 2. Prerequisites.- 3. Membrane Systems with Symbol-Objects.- 4. Trading Evolution for Communication.- 5. Structuring Objects.- 6. Networks of Membranes.- 7. Trading Space for Time.- 8. Further Technical Results.- 9. (Attempts to Get) Back to Reality.- Open Problems.- Universality Results. Bibliography.- Index.
1,760 citations
•
30 Nov 1993
TL;DR: Computational complexity is the realm of mathematical models and techniques for establishing impossibility proofs for proving formally that there can be no algorithm for the given problem which runs faster than the current one.
Abstract: Once we have developed an algorithm (q.v.) for solving a computational problem and analyzed its worst-case time requirements as a function of the size of its input (most usefully, in terms of the O-notation; see ALGORITHMS, ANALYSIS OF), it is inevitable to ask the question: "Can we do better?" In a typical problem, we may be able to devise new algorithms for the problem that are more and more efficient. But eventually, this line of research often seems to hit an invisible barrier, a level beyond whch improvements are very difficult, seemingly impossible, to come by. After many unsuccessful attempts, algorithm designers inevitably start to wonder if there is something inherent in the problem that makes it impossible to devise algorithms that are faster than the current one. They may try to develop mathematical techniques for proving formally that there can be no algorithm for the given problem which runs faster than the current one. Such a proof would be valuable, as it would suggest that it is futile to keep working on improved algorithms for this problem, that further improvements are certainly impossible. The realm of mathematical models and techniques for establishing such impossibility proofs is called computational complexity.
965 citations
Additional excerpts
...STCON is also known as PATH [17] and REACHABILITY [12]....
[...]
••
TL;DR: A super-polynomial lower bound is given for the size of circuits of fixed depth computing the parity function and connections are given to the theory of programmable logic arrays and to the relativization of the polynomial-time hierarchy.
Abstract: A super-polynomial lower bound is given for the size of circuits of fixed depth computing the parity function. Introducing the notion of polynomial-size, constant-depth reduction, similar results are shown for the majority, multiplication, and transitive closure functions. Connections are given to the theory of programmable logic arrays and to the relativization of the polynomial-time hierarchy.
915 citations