Disjunctive Normal Form for Multi-Agent Modal Logics Based on Logical Separability

Total Page:16

File Type:pdf, Size:1020Kb

Disjunctive Normal Form for Multi-Agent Modal Logics Based on Logical Separability The Thirty-Third AAAI Conference on Artificial Intelligence (AAAI-19) Disjunctive Normal Form for Multi-Agent Modal Logics Based on Logical Separability Liangda Fang,1 Kewen Wang,2 Zhe Wang,2 Ximing Wen3 1Department of Computer Science, Jinan University, China 2School of Information and Communication Technology, Griffith University, Australia 3Guangdong Institute of Public Administration, Guangzhou, China [email protected], fk.wang,zhe.wangg@griffith.edu.au, [email protected] Abstract and the action preconditions. If a normal form is defined such that its entailment check, bounded conjunction and for- Modal logics are primary formalisms for multi-agent systems getting are tractable, then both progression and entailment but major reasoning tasks in such logics are intractable, which check in the normal form are tractable (Bienvenu, Fargier, impedes applications of multi-agent modal logics such as au- tomatic planning. One technique of tackling the intractabil- and Marquis 2010). As a result, in the normal form, an ef- ity is to identify a fragment called a normal form of multi- fective search algorithm is obtained. In the case of proposi- agent logics such that it is expressive but tractable for reason- tional logic, a normal form can be disjunctive normal form ing tasks such as entailment checking, bounded conjunction (DNF), conjunctive normal form (CNF) or prime implicate. transformation and forgetting. For instance, DNF of proposi- Usually, a planner based on DNF is faster than those based tional logic is tractable for these reasoning tasks. In this pa- on CNF and prime implicate (To, Pontelli, and Son 2011; per, we first introduce a notion of logical separability and then To, Son, and Pontelli 2011). This is due to the fact that, define a novel disjunctive normal form SDNF for the multi- in propositional logic, DNF possesses tractable entailment agent logic Kn, which overcomes some shortcomings of exist- check, bounded conjunction and forgetting, but CNF and ing approaches. In particular, we show that every modal for- prime implicate do not (Darwiche and Marquis 2002). mula in Kn can be equivalently casted as a formula in SDNF, major reasoning tasks tractable in propositional DNF are also Thus, in order to develop effective search algorithms to tractable in SDNF, and moreover, formulas in SDNF enjoy multi-agent epistemic planning, researchers aimed to de- the property of logical separability. To demonstrate the use- velop suitable DNF-like normal forms for multi-agent modal fulness of our approach, we apply SDNF in multi-agent epis- logics such that major reasoning tasks, such as bounded temic planning. Finally, we extend these results to three more conjunction, forgetting and entailment check, are tractable. complex multi-agent logics Dn, K45n and KD45n. Such a disjunctive normal form, named S5-DNF, is defined for modal logic S5 but only for the single-agent case (Bi- 1 Introduction envenu, Fargier, and Marquis 2010). When this DNF is ex- tended to multi-agent modal logics, both entailment and for- It is crucial for an intelligent agent system to be capable getting will not be tractable any longer. Two other normal of representing and reasoning about high-order knowledge forms, cover disjunctive normal forms (CDNFs) (ten Cate et in the multi-agent setting. A general representative frame- al. 2006) and prime implicate normal forms (PINFs) (Bi- work for these scenarios is multi-agent modal logics. How- envenu 2008), have been introduced for description logic ever, some important reasoning tasks, including satisfiabil- ALC, a syntactic variant of the multi-agent modal logic ity checking and forgetting, are intractable in such logics Kn. However, these two normal forms have some shortcom- (Halpern and Moses 1992; Bienvenu 2009). The intractabil- ings. CDNF is relatively less compact, that is, the CDNF ity results impede applications of multi-agent modal logics, representation is exponential large for some simple formu- e.g., multi-agent epistemic planning (Kominis and Geffner las. Bounded conjunction for PINF is intractable and in 2015; Huang et al. 2017). the worst case a compiled formula in PINF is double ex- Traditionally, the forward search algorithm is effective for ponentially large. In addition, Moss (2007) introduced a agent planning (Bryce, Kambhampati, and Smith 2006), in canonical form of modal formulas for Kn and Dn. While which the search is performed in the space of knowledge modal formulas can be equivalently transformed into Moss’ bases (KBs) proceeding forward from the initial KB towards canonical form, the complexity of the transformation is non- a goal KB entailing the goal formula. Two types of reasoning elementary. Hence, Moss’ canonical formulas are not a prac- tasks are essential to the search algorithm, namely, progres- tical normal form for epistemic planning. sion and entailment check. The progression updates KBs ac- logic sepa- cording to the action effects while the entailment check is In this paper, we first formulate the notion of rability needed to decide if the current KB entails the goal formula in multi-agent modal logic, which was originally in- vestigated for first-order logic (Levesque 1998). Informally, Copyright c 2019, Association for the Advancement of Artificial a conjunction φ of formulas is logically separable w.r.t. a Intelligence (www.aaai.org). All rights reserved. form of reasoning if the reasoning for φ can be reduced to 2817 the same reasoning for the conjuncts of φ. For example, the P (φ) for the set of variables appearing in φ. A formula φ is formula φ = (p ! q) ^ (q ! r) is not logically separable smaller than , if jφj < j j. since it logically implies a conjunct p ! r that is not de- The notions of propositional literals, terms (TE), clauses rived by any single conjunct of φ. By conjoining φ with the (CL), disjunctive normal forms (DNF) and conjunctive nor- implicit conjunct, the new formula becomes logically sepa- mal forms (CNF) are defined as usual. rable. A modal literal is a formula of the form iφ or ♦iφ, where Based on the logical separability, we introduce a disjunc- φ is a modal formula. A modal literal is positive (resp. neg- tive normal form, called separability-based DNF (or SDNF) ative), if it is of the form iφ (resp. ♦iφ). A formula is ba- for multi-agent modal logics Kn, Dn, K45n and KD45n. sic, if it is a propositional formula or modal literal. A modal Similarly, separability-based CNF (or SCNF) can also be term (resp. clause) is a conjunction (resp. disjunction) of defined. For these two normal forms and the two existing basic formulas. For a modal term (resp. clause) φ, we use normal forms for multi-agent modal logics, we investigate Prop(φ) for the set of propositional components (i.e., maxi- the expressiveness, succinctness, queries and transforma- mal propositional subformulas) that are conjuncts (resp. dis- tions (Darwiche and Marquis 2002). Thus we obtain an al- juncts) of φ, Bi(φ) for the set of formulas such that i most complete knowledge compilation map for multi-agent is a conjunct (resp. disjunct) of φ, and Di(φ) for the set of modal logics. formulas such that ♦i is a conjunct (resp. disjunct) of The main contributions of this paper contain: φ. For example, consider the modal term φ = p ^ :q ^ 1. We formulate the concept of logical separability for i:p^iq ^♦i(:p_q)^♦ip. Thus, Prop(φ) = fp ^ :qg, modal terms and formalise some desirable properties for Bi(φ) = f:p; qg, and Di(φ) = f:p _ q; pg. them. Thanks to the notion of logical separability, we are Definition 2.1. A Kripke model M is a tuple hS;R;Vi where able to define separability-based DNF and CNF, two novel normal forms for Kn. • S is a non-empty set of possible worlds; 2. For the two new normal forms as well as CDNF • R = fRi j i 2 Ag where Ri is a binary relation on S; and PINF, we investigate the expressiveness, succinctness, • V is a function assigning to each s 2 S a subset of P . queries and transformations. To the best of our knowledge, we are the first to construct this map for multi-agent modal A pointed Kripke model is a pair (M; s), where M is a Kripke model and s is a world of M, called the actual world. logics. Interestingly, SDNFL0 possesses all properties that propositional DNF has, e.g., polytime satisfiability, bounded For convenience, we assume that Kripke models are conjunction, forgetting and so on. In this sense, SDNF is (M; s) L pointed. Given a Kripke model and an -formula a proper generalisation of propositional DNF for the multi- φ, the satisfaction relation M; s j= φ is defined as usual and agent modal logic Kn. in this case, we say that (M; s) is a model of φ. A modal 3. As a case study, we illustrate the application of our re- formula φ is satisfiable if it has a model; φ entails , written sults in multi-agent epistemic planning. φ j= , if every model of φ is also a model of ; φ and 4. We extend the results of knowledge compilation for are equivalent, written φ ≡ , if φ j= and j= φ. 0 Kn to Dn without any modification on normal forms, and Throughout this paper, we use L and L for fragments of 0 to K45n and KD45n under the condition that no consecutive L L L , and 0 and 0 for propositional fragments. We also modalities of the same agent appear in the given formula. assume that every propositional term and clause has a poly- 0 nomial representation in L0 and L0. All propositional frag- 2 The multi-agent modal logic Kn ments considered in (Darwiche and Marquis 2002) conform with this assumption except for full DNF, that is, if each of In this section, we first recall the syntax and semantics of its variables appears exactly once in every term.
Recommended publications
  • Computational Complexity: a Modern Approach
    i Computational Complexity: A Modern Approach Draft of a book: Dated January 2007 Comments welcome! Sanjeev Arora and Boaz Barak Princeton University [email protected] Not to be reproduced or distributed without the authors’ permission This is an Internet draft. Some chapters are more finished than others. References and attributions are very preliminary and we apologize in advance for any omissions (but hope you will nevertheless point them out to us). Please send us bugs, typos, missing references or general comments to [email protected] — Thank You!! DRAFT ii DRAFT Chapter 9 Complexity of counting “It is an empirical fact that for many combinatorial problems the detection of the existence of a solution is easy, yet no computationally efficient method is known for counting their number.... for a variety of problems this phenomenon can be explained.” L. Valiant 1979 The class NP captures the difficulty of finding certificates. However, in many contexts, one is interested not just in a single certificate, but actually counting the number of certificates. This chapter studies #P, (pronounced “sharp p”), a complexity class that captures this notion. Counting problems arise in diverse fields, often in situations having to do with estimations of probability. Examples include statistical estimation, statistical physics, network design, and more. Counting problems are also studied in a field of mathematics called enumerative combinatorics, which tries to obtain closed-form mathematical expressions for counting problems. To give an example, in the 19th century Kirchoff showed how to count the number of spanning trees in a graph using a simple determinant computation. Results in this chapter will show that for many natural counting problems, such efficiently computable expressions are unlikely to exist.
    [Show full text]
  • Lecture 10: Space Complexity III
    Space Complexity Classes: NL and L Reductions NL-completeness The Relation between NL and coNL A Relation Among the Complexity Classes Lecture 10: Space Complexity III Arijit Bishnu 27.03.2010 Space Complexity Classes: NL and L Reductions NL-completeness The Relation between NL and coNL A Relation Among the Complexity Classes Outline 1 Space Complexity Classes: NL and L 2 Reductions 3 NL-completeness 4 The Relation between NL and coNL 5 A Relation Among the Complexity Classes Space Complexity Classes: NL and L Reductions NL-completeness The Relation between NL and coNL A Relation Among the Complexity Classes Outline 1 Space Complexity Classes: NL and L 2 Reductions 3 NL-completeness 4 The Relation between NL and coNL 5 A Relation Among the Complexity Classes Definition for Recapitulation S c NPSPACE = c>0 NSPACE(n ). The class NPSPACE is an analog of the class NP. Definition L = SPACE(log n). Definition NL = NSPACE(log n). Space Complexity Classes: NL and L Reductions NL-completeness The Relation between NL and coNL A Relation Among the Complexity Classes Space Complexity Classes Definition for Recapitulation S c PSPACE = c>0 SPACE(n ). The class PSPACE is an analog of the class P. Definition L = SPACE(log n). Definition NL = NSPACE(log n). Space Complexity Classes: NL and L Reductions NL-completeness The Relation between NL and coNL A Relation Among the Complexity Classes Space Complexity Classes Definition for Recapitulation S c PSPACE = c>0 SPACE(n ). The class PSPACE is an analog of the class P. Definition for Recapitulation S c NPSPACE = c>0 NSPACE(n ).
    [Show full text]
  • Computational Complexity: a Modern Approach
    i Computational Complexity: A Modern Approach Draft of a book: Dated January 2007 Comments welcome! Sanjeev Arora and Boaz Barak Princeton University [email protected] Not to be reproduced or distributed without the authors’ permission This is an Internet draft. Some chapters are more finished than others. References and attributions are very preliminary and we apologize in advance for any omissions (but hope you will nevertheless point them out to us). Please send us bugs, typos, missing references or general comments to [email protected] — Thank You!! DRAFT ii DRAFT Chapter 5 The Polynomial Hierarchy and Alternations “..synthesizing circuits is exceedingly difficulty. It is even more difficult to show that a circuit found in this way is the most economical one to realize a function. The difficulty springs from the large number of essentially different networks available.” Claude Shannon 1949 This chapter discusses the polynomial hierarchy, a generalization of P, NP and coNP that tends to crop up in many complexity theoretic investigations (including several chapters of this book). We will provide three equivalent definitions for the polynomial hierarchy, using quantified predicates, alternating Turing machines, and oracle TMs (a fourth definition, using uniform families of circuits, will be given in Chapter 6). We also use the hierarchy to show that solving the SAT problem requires either linear space or super-linear time. p p 5.1 The classes Σ2 and Π2 To understand the need for going beyond nondeterminism, let’s recall an NP problem, INDSET, for which we do have a short certificate of membership: INDSET = {hG, ki : graph G has an independent set of size ≥ k} .
    [Show full text]
  • 1 the Class Conp
    80240233: Computational Complexity Lecture 3 ITCS, Tsinghua Univesity, Fall 2007 16 October 2007 Instructor: Andrej Bogdanov Notes by: Jialin Zhang and Pinyan Lu In this lecture, we introduce the complexity class coNP, the Polynomial Hierarchy and the notion of oracle. 1 The class coNP The complexity class NP contains decision problems asking this kind of question: On input x, ∃y : |y| = p(|x|) s.t.V (x, y), where p is a polynomial and V is a relation which can be computed by a polynomial time Turing Machine (TM). Some examples: SAT: given boolean formula φ, does φ have a satisfiable assignment? MATCHING: given a graph G, does it have a perfect matching? Now, consider the opposite problems of SAT and MATCHING. UNSAT: given boolean formula φ, does φ have no satisfiable assignment? UNMATCHING: given a graph G, does it have no perfect matching? This kind of problems are called coNP problems. Formally we have the following definition. Definition 1. For every L ⊆ {0, 1}∗, we say that L ∈ coNP if and only if L ∈ NP, i.e. L ∈ coNP iff there exist a polynomial-time TM A and a polynomial p, such that x ∈ L ⇔ ∀y, |y| = p(|x|) A(x, y) accepts. Then the natural question is how P, NP and coNP relate. First we have the following trivial relationships. Theorem 2. P ⊆ NP, P ⊆ coNP For instance, MATCHING ∈ coNP. And we already know that SAT ∈ NP. Is SAT ∈ coNP? If it is true, then we will have NP = coNP. This is the following theorem. Theorem 3. If SAT ∈ coNP, then NP = coNP.
    [Show full text]
  • Descriptive Complexity Theories
    Descriptive Complexity Theories Joerg FLUM ABSTRACT: In this article we review some of the main results of descriptive complexity theory in order to make the reader familiar with the nature of the investigations in this area. We start by presenting the characterization of automata recognizable languages by monadic second-order logic. Afterwards we explain the characteri- zation of various logics by fixed-point logics. We assume familiarity with logic but try to keep knowledge of complexity theory to a minimum. Keywords: Computational complexity theory, complexity classes, descriptive characterizations, monadic second-order logic, fixed-point logic, Turing machine. Complexity theory or more precisely, computational complexity theory (cf. Papadimitriou 1994), tries to classify problems according to the complexity of algorithms solving them. Of course, we can think of various ways of measuring the complexity of an al- gorithm, but the most important and relevant ones are time and space. That is, we think we have a type of machine capable, in principle, to carry out any algorithm, a so-called general purpose machine (or, general purpose computer), and we measure the com- plexity of an algorithm in terms of the time or the number of steps needed to carry out this algorithm. By space, we mean the amount of memory the algorithm uses. Time bounds yield (time) complexity classes consisting of all problems solvable by an algorithm keeping to the time bound. Similarly, space complexity classes are obtained. It turns out that these definitions are quite robust in the sense that, for reasonable time or space bounds, the corresponding complexity classes do not depend on the special type of machine model chosen.
    [Show full text]
  • Descriptive Complexity: a Logicians Approach to Computation
    Descriptive Complexity: A Logician’s Approach to Computation N. Immerman basic issue in computer science is there are exponentially many possibilities, the complexity of problems. If one is but in this case no known algorithm is faster doing a calculation once on a than exponential time. medium-sized input, the simplest al- Computational complexity measures how gorithm may be the best method to much time and/or memory space is needed as Ause, even if it is not the fastest. However, when a function of the input size. Let TIME[t(n)] be the one has a subproblem that will have to be solved set of problems that can be solved by algorithms millions of times, optimization is important. A that perform at most O(t(n)) steps for inputs of fundamental issue in theoretical computer sci- size n. The complexity class polynomial time (P) ence is the computational complexity of prob- is the set of problems that are solvable in time lems. How much time and how much memory at most some polynomial in n. space is needed to solve a particular problem? ∞ Here are a few examples of such problems: P= TIME[nk] 1. Reachability: Given a directed graph and k[=1 two specified points s,t, determine if there Even though the class TIME[t(n)] is sensitive is a path from s to t. A simple, linear-time to the exact machine model used in the com- algorithm marks s and then continues to putations, the class P is quite robust. The prob- mark every vertex at the head of an edge lems Reachability and Min-triangulation are el- whose tail is marked.
    [Show full text]
  • Complexity Theory - Tutorial
    Complexity Theory - Tutorial Ivan Gavran January 31st, 2017 Part 1 1. Decide whether the following statements are true or false and justify your decision. (a) For every class C and A 2 C; if A is C-complete under Karp reduc- tions, then C ⊂ P A Solution True. Any B 2 C can be transformed into instance of A. But this - by the definition - means that B can be decided by polynomial computation (do the transformation into an instance of A) and then asking oracle A. (b) If L is NP-complete and L0 is coNP complete, then L \ L0 is NP \ coNP complete. Solution False. SAT is NP-complete. NOSAT is coNP-complete. Their intersection is empty (so that language would trivially always return false, but we know that there are non-trivial languages in NP \ coNP , primality, for example.). (c) If PH=PSPACE, then PH collapses to some level Σk Solution True. There is a complete problem for PSPACE (TQBF, for example). Since PH = PSPACE, this problem is in PH and therefore it is in Σk, for some k. But now all the problems in PH can be in polynomial time reduced to that one. Therefore, PH collapses to Σk. (d) There is an undecidable language in P=poly Solution True All unary languages are in P/poly, unary variant of halting problem is in P/poly. (e) If P=NP, then NP=coNP. Solution True Since NP = P and P ⊂ coNP it is clear that NP ⊂ coNP . Take L 2 coNP . This means that L¯ 2 NP = P .
    [Show full text]
  • 6.080 / 6.089 Great Ideas in Theoretical Computer Science Spring 2008
    MIT OpenCourseWare http://ocw.mit.edu 6.080 / 6.089 Great Ideas in Theoretical Computer Science Spring 2008 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms. 6.080/6.089 GITCS DATE Lecture 12 Lecturer: Scott Aaronson Scribe: Mergen Nachin 1 Review of last lecture • NP-completeness in practice. We discussed many of the approaches people use to cope with NP-complete problems in real life. These include brute-force search (for small instance sizes), cleverly-optimized backtrack search, fixed-parameter algorithms, approximation algorithms, and heuristic algorithms like simulated annealing (which don’t always work but often do pretty well in practice). • Factoring, as an intermediate problem between P and NP-complete. We saw how factoring, though believed to be hard for classical computers, has a special structure that makes it different from any known NP-complete problem. • coNP. 2 Space complexity Now we will categorize problems in terms of how much memory they use. Definition 1 L is in PSPACE if there exists a poly-space Turing machine M such that for all x, x is in L if and only if M(x) accepts. Just like with NP, we can define PSPACE-hard and PSPACE-complete problems. An interesting example of a PSPACE-complete problem is n-by-n chess: Given an arrangement of pieces on an n-by-n chessboard, and assuming a polynomial upper bound on the number of moves, decide whether White has a winning strategy. (Note that we need to generalize chess to an n-by-n board, since standard 8-by-8 chess is a finite game that’s completely solvable in O(1) time.) Let’s now define another complexity class called EXP, which is apparently even bigger than PSPACE.
    [Show full text]
  • Complexity Classes
    CHAPTER Complexity Classes In an ideal world, each computational problem would be classified at least approximately by its use of computational resources. Unfortunately, our ability to so classify some important prob- lems is limited. We must be content to show that such problems fall into general complexity classes, such as the polynomial-time problems P, problems whose running time on a determin- istic Turing machine is a polynomial in the length of its input, or NP, the polynomial-time problems on nondeterministic Turing machines. Many complexity classes contain “complete problems,” problems that are hardest in the class. If the complexity of one complete problem is known, that of all complete problems is known. Thus, it is very useful to know that a problem is complete for a particular complexity class. For example, the class of NP-complete problems, the hardest problems in NP, contains many hundreds of important combinatorial problems such as the Traveling Salesperson Prob- lem. It is known that each NP-complete problem can be solved in time exponential in the size of the problem, but it is not known whether they can be solved in polynomial time. Whether ? P and NP are equal or not is known as the P = NP question. Decades of research have been devoted to this question without success. As a consequence, knowing that a problem is NP- complete is good evidence that it is an exponential-time problem. On the other hand, if one such problem were shown to be in P, all such problems would be been shown to be in P,a result that would be most important.
    [Show full text]
  • Notes for Lecture 3 1 Space-Bounded Complexity Classes
    U.C. Berkeley — CS278: Computational Complexity Handout N3 Professor Luca Trevisan 9/6/2004 Notes for Lecture 3 Revised 10/6/04 1 Space-Bounded Complexity Classes A machine solves a problem using space s(·) if, for every input x, the machine outputs the correct answer and uses only the first s(|x|) cells of the tape. For a standard Turing machine, we can’t do better than linear space since x itself must be on the tape. So we will often consider a machine with multiple tapes: a read-only “input” tape, a read/write “work” or “memory” tape, and possibly a write-once “output” tape. Then we can say the machine uses space s if for input x, it uses only the first s(|x|) cells of the work tape. We denote by L the set of decision problems solvable in O(log n) space. We denote by PSPACE the set of decision problems solvable in polynomial space. A first observation is that a space-efficient machine is, to a certain extent, also a time-efficient one. In general we denote by SPACE(s(n)) the set of decision problems that can be solved using space at most s(n) on inputs of length n. Theorem 1 If a machine always halts, and uses s(·) space, with s(n) ≥ log n, then it runs in time 2O(s(n)). Proof: Call the “configuration” of a machine M on input x a description of the state of M, the position of the input tape, and the contents of the work tape at a given time.
    [Show full text]
  • Complexity Classes De Ned by Counting Quanti Ers*
    Complexity classes dened by counting quantiers Jacob o Toran Dept LSI Universidad Politecnica de Cataluna Pau Gargallo Barcelona Spain Abstract We study the p olynomial time counting hierarchy a hierarchy of complexity classes related to the notion of counting We investigate some of their structural prop erties settling many op en questions dealing with oracle characterizations closure under b o olean op erations and relations with other complexity classes We develop a new combinatorial technique to obtain relativized separations for some of the studied classes which imply absolute separations for some logarithmic time b ounded complexity classes Intro duction One of the main goals of complexity theory is the classication of computational problems in complexity classes according to the amount of resources these problems need Probably the b est known complexity classes are P and NP since b oth of them capture the complexity of many natural problems and also b ecause the long standing op en question P NP has motivated most of the research in the area The similarities in the denitions of the class NP and the recursiontheoretic class of the recursively enumerable sets b oth can b e characterized by an existential quantier provoked the translation of other recursion theoretic notions into the eld of complexity theory and the analogous concept to the arithmetic hierarchy the p olynomial time hierarchy St was dened The idea is a natural generalization of the class NP and provided a go o d to ol to classify more complex problems It was
    [Show full text]
  • 6.845 Quantum Complexity Theory, Lecture 19
    6.896 Quantum Complexity Theory November 6, 2008 Lecture 19 Lecturer: Scott Aaronson Scribe: Joshua Horowitz On Election Day, November 4, 2008, the people voted and their decision was clear: Prof. Aaronson would talk about quantum computing with closed time-like curves. But before we move into strange physics with even stranger complexity-theoretic implications, we need to fill in a more basic gap in our discussion so far – quantum space-complexity classes. 1 BQPSPACE Recall that PSPACE is the class of all decision problems which can be solved by some classical Turing machine, whose tape-length is bounded by some polynomial of the input-length. To make BQPSPACE, we quantize this in the most straightforward way possible. That is, we let BQPSPACE be the class of all decision problems which can be solved with a bounded rate of error (the B) by some quantum Turing machine (the Q), whose tape-length is bounded by some polynomial of the input-length (the PSPACE). As we’ve seen before, allowing quantum computers to simulate classical ones gives the relationship PSPACE ⊆ BQPSPACE. If we believe that BQP is larger than P, we might suspect by analogy that BQPSPACE is larger than PSPACE. But it turns out that the analogy isn’t such a great one, since space and time seem to work in very different ways. As an example of the failure of the time-space analogy, take PSPACE vs. NPSPACE. We certainly believe that nondeterminism gives classical machines exponential time speed-ups for some problems in NP, making P �= NP at least in our minds.
    [Show full text]