Message Board

Dear readers, authors and reviewers,you can add a message on this page. We will reply to you as soon as possible!

2021 Volume 43 Issue 10
Article Contents

TIAN Keming, WU Jianrong. Application of Fuzzy Quasi-Metric in Complexity Analysis[J]. Journal of Southwest University Natural Science Edition, 2021, 43(10): 110-116. doi: 10.13718/j.cnki.xdzk.2021.10.015
Citation: TIAN Keming, WU Jianrong. Application of Fuzzy Quasi-Metric in Complexity Analysis[J]. Journal of Southwest University Natural Science Edition, 2021, 43(10): 110-116. doi: 10.13718/j.cnki.xdzk.2021.10.015

Application of Fuzzy Quasi-Metric in Complexity Analysis

More Information
  • Corresponding author: WU Jianrong
  • Received Date: 22/09/2020
    Available Online: 20/10/2021
  • MSC: O159

  • In complexity theory, complexity analysis is mainly aimed at the efficiency of the algorithm. Generally, in complexity analysis, more attention is paid to the asymptotic efficiency of the algorithm. In recent years, the application of quasi-metric in complexity analysis has been widely studied by scholars, but it also has some limitations. For example, quasi-metric is not suitable for describing the asymptotic efficiency of algorithms. In order to solve this problem, a fuzzy quasi-metric in a complexity space is introduced and used to describe the asymptotic efficiency of algorithms. After giving some results of the fuzzy quasi-metric, a fixed point theorem in a complexity space is established. As its application, the existence and uniqueness of the solution of recurrence equations associated with Divide-and-Conquer algorithm and the existence and uniqueness of the solution of recurrence equations associated with Quicksort algorithm are proven. The above results construct the relationship between fuzzy quasi-metric and the asymptotic efficiency of the algorithm, and provide a new and effective approach for the further research of application of fuzzy quasi-metric in algorithms.
  • 加载中
  • [1] SCHELLEKENS M. The Smyth Completion: a Common Foundation for Denotational Semantics and Complexity Analysis[J]. Electronic Notes in Theoretical Computer Science, 1995, 1: 535-556. doi: 10.1016/S1571-0661(04)00029-5

    CrossRef Google Scholar

    [2] ROMAGUERA S, SCHELLEKENS M. Quasi-Metric Properties of Complexity Spaces[J]. Topology and Its Applications, 1999, 98(1-3): 311-322. doi: 10.1016/S0166-8641(98)00102-3

    CrossRef Google Scholar

    [3] ROMAGUERA S, SCHELLEKENS M. Duality and Quasi-Normability for Complexity Spaces[J]. Applied General Topology, 2002, 3(1): 91-112. doi: 10.4995/agt.2002.2116

    CrossRef Google Scholar

    [4] GARCÍA-RAFFI L M, ROMAGUERA S, SÁNCHEZ-PÉREZ E A. Sequence Spaces and Asymmetric Norms in the Theory of Computational Complexity[J]. Mathematical and Computer Modelling, 2002, 36(1-2): 1-11. doi: 10.1016/S0895-7177(02)00100-0

    CrossRef Google Scholar

    [5] ROMAGUERA S, SÁNCHEZ-PÉREZ E A, VALERO O. The Dual Complexity Space as the Dual of a Normed Cone[J]. Electronic Notes in Theoretical Computer Science, 2006, 161(2): 165-174.

    Google Scholar

    [6] ROMAGUERA S, SCHELLEKENS M. Partial Metric Monoids and Semivaluation Spaces[J]. Topology and Its Applications, 2005, 153(5-6): 948-962. doi: 10.1016/j.topol.2005.01.023

    CrossRef Google Scholar

    [7] SCHWEIZER B, SKLAR A. Statistical Metric Spaces[J]. Pacific Journal of Mathematics, 1960, 10(1): 314-334.

    Google Scholar

    [8] GREGORI V, ROMAGUERA S. Fuzzy Quasi-Metric Spaces[J]. Applied General Topology, 2004, 5(1): 129-136. doi: 10.4995/agt.2004.2001

    CrossRef Google Scholar

    [9] ROMAGUERA S, TIRADO P. The Complexity Probabilistic Quasi- Metric Space[J]. Journal of Mathematical Analysis and Applications, 2011, 376(2): 732-740. doi: 10.1016/j.jmaa.2010.11.056

    CrossRef Google Scholar

    [10] KRAMOSIL I, MICHALEK J. Fuzzy Metric and Statistical Metric Spaces[J]. Kyberneticall Praha, 1975, 11(5): 215-229.

    Google Scholar

    [11] 杨浩, 吴健荣. 模糊度量空间中的有界集[J]. 西南大学学报(自然科学版), 2019, 41(10): 45-50.

    Google Scholar

    [12] 杨浩, 吴健荣. 模糊度量空间中的伪度量结构及等距同构[J]. 西南大学学报(自然科学版), 2021, 43(6): 95-100.

    Google Scholar

    [13] REILLY I L, SUBRAHMANYAM P V, VAMANAMURTHY M K. Cauchy Sequences in Quasi-Pseudo-Metric Spaces[J]. Monatshefte für Mathematik, 1982, 93(2): 127-140. doi: 10.1007/BF01301400

    CrossRef Google Scholar

    [14] AHO A V, HOPCROFT J E, ULLMAN J. Data Structures and Algorithms[M]. New York: Addison-Wesley, 1983.

    Google Scholar

    [15] KRUSE R L. Data Structures and Program Design[M]. Englewood: Prentice-Hall, 1987.

    Google Scholar

    [16] ROMAGUERA S, SAPENA A, TIRADO P. The Banach Fixed Point Theorem in Fuzzy Quasi-Metric Spaces with Application to the Domain of Words[J]. Topology and Its Applications, 2007, 154(10): 2196-2203. doi: 10.1016/j.topol.2006.09.018

    CrossRef Google Scholar

    [17] SAADATI R, VAEZPOUR S M, CHO Y J. Quicksort Algorithm: Application of a Fixed Point Theorem in Intuitionistic Fuzzy Quasi-Metric Spaces at a Domain of Words[J]. Journal of Computational and Applied Mathematics, 2009, 228(1): 219-225. doi: 10.1016/j.cam.2008.09.013

    CrossRef Google Scholar

  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Article Metrics

Article views(757) PDF downloads(221) Cited by(0)

Access History

Other Articles By Authors

Application of Fuzzy Quasi-Metric in Complexity Analysis

    Corresponding author: WU Jianrong

Abstract: In complexity theory, complexity analysis is mainly aimed at the efficiency of the algorithm. Generally, in complexity analysis, more attention is paid to the asymptotic efficiency of the algorithm. In recent years, the application of quasi-metric in complexity analysis has been widely studied by scholars, but it also has some limitations. For example, quasi-metric is not suitable for describing the asymptotic efficiency of algorithms. In order to solve this problem, a fuzzy quasi-metric in a complexity space is introduced and used to describe the asymptotic efficiency of algorithms. After giving some results of the fuzzy quasi-metric, a fixed point theorem in a complexity space is established. As its application, the existence and uniqueness of the solution of recurrence equations associated with Divide-and-Conquer algorithm and the existence and uniqueness of the solution of recurrence equations associated with Quicksort algorithm are proven. The above results construct the relationship between fuzzy quasi-metric and the asymptotic efficiency of the algorithm, and provide a new and effective approach for the further research of application of fuzzy quasi-metric in algorithms.

  • 开放科学(资源服务)标志码(OSID):

  • 文献[1]为了构建复杂性分析的拓扑基础,在复杂性空间中引入了一种拟度量,并将其应用于分治算法的效率分析中. 随后,文献[2]进一步研究了该拟度量空间的性质,证明了复杂性空间是Smyth完备的,可以被建模为拟范数半线性空间. 关于此空间的更多结果详见文献[3-6].

    算法的渐近效率在复杂性分析中具有重要的应用. 然而,正如本文例2所述,文献[1]中引入的拟度量并不适合刻画算法的渐近效率.

    为解决上述问题,在本文中,我们在复杂性函数集上引入了一种模糊拟度量,并且证明了它可以刻画算法的渐近效率. 此外,我们建立了一个不动点定理,并应用此定理证明了与分治算法和快速排序算法相关的递归方程解的存在性和唯一性.

1.   预备知识
  • 在本文中,符号${{\mathbb{N}}_{+}}$表示非负整数集.

    定义1[7]  设*:[0, 1]×[0, 1]→[0, 1]是一个二元运算,如果

    (a) *满足结合律和交换律;

    (b) *是连续的;

    (c) ∀a∈[0, 1],a*1=a

    (d) 当acbd(abcd∈[0, 1])时,a*bc*d.

    则称*是一个连续t-模.

    注1  当ab∈[0, 1]时,对任意的连续t-模*,aba*b恒成立.

    定义2[8]  设X是一个非空集合,*是一个连续t-模,MX×X×[0,∞)上的模糊集,对∀xyzX,有

    (FQM1) M(xy,0)=0;

    (FQM2) ∀t>0,M(xyt)=M(yxt)=1$\Leftrightarrow $x=y

    (FQM3) ∀ts≥0,M(xzt+s)≥M(xyt)*M(yzs);

    (FQM4) M(xy,·):[0,∞)→[0, 1]是左连续的.

    则称(M,*)为X上的一个模糊拟度量.

    注2  文献[9]介绍的概率拟度量(PC,∧)可看作一个特殊的模糊拟度量.

    注3M还满足

    (FQM5) ∀t>0,M(xyt)=M(yxt).

    则(M,*)为文献[10]意义下的一个模糊度量,有关模糊度量的更多结果详见文献[11-12].

    设(M,*)是X上的模糊拟度量,若M-1X×X×[0,∞)上的函数,且满足M-1(xyt)=M(yxt),则(M-1,*)也是X上的一个模糊拟度量. 此外,若MiX×X×[0,∞)上的函数,且满足Mi(xyt)=min{M(xyt),M-1(xyt)},则(Mi,*)是X上的一个模糊度量.

    文献[10]证明了:X上任意的模糊度量Mi可以诱导生成一个T0拓扑τMi,该拓扑有一个开球族$\left\{B_{M^{i}}\left(x, \frac{1}{n}, \frac{1}{n}\right): n \in \mathbb{N}_{+}\right\}$的局部基,其中

    此外τMiT2分离的和第一可数的.

    定义3[8]  设{xn}是模糊拟度量空间(XM,*)中的序列,若对∀ε∈(0,1),t>0,都存在n0${{\mathbb{N}}_{+}}$,使得当mnn0M(xnxmt)>1-ε,则称序列{xn}是左K-柯西列.

    定义4  设(XM,*)是一个模糊拟度量空间,若对X中任意的左K-柯西列{xn},存在xX,使得$\lim\limits _{n \rightarrow \infty} M\left(x, x_{n}, t\right)=\lim \limits_{n \rightarrow \infty} M\left(x_{n}, x, t\right)=1$,则称(XM,*)是Smyth完备的.

    例1[8]  设(Xd)是一个拟度量空间,Md是定义在X×X×[0,∞)上的函数,且对任意的xyXt≥0,有

    若*为任意的连续t-模,则(Md,*)为X上的一个模糊拟度量,称(Md,*)是由d导出的标准模糊拟度量.

    接下来我们介绍拟度量空间的一些基本概念.

    定义5[13]  设(Xd)是一个拟度量空间,若对∀ε∈(0,1),t>0,都存在n0${{\mathbb{N}}_{+}}$,当mnn0d(xnxm)<ε,则称序列{xn}为左K-柯西列.

    定义6[13]  设(Xd)是一个拟度量空间,若(Xd)中任意的左K-柯西列是收敛的,则称(Xd)是Smyth完备的.

    定理1[2]  设(Xd)是一个拟度量空间,(Xd)是Smyth完备的当且仅当(Xd)中任意的左K-柯西列是收敛的,即存在xX使得$\lim \limits_{n \rightarrow \infty} d^{s}\left(x, x_{n}\right)=0$.

    定义7[1]  设$C=\left\{f: \mathbb{N}_{+} \longrightarrow(0,+\infty]: \sum\limits_{n=1}^{+\infty} 2^{-n} \frac{1}{f(n)} <+\infty\right\} $,对∀fgC,定义

    则称C为复杂性函数集,称dC为复杂性拟度量,称(CdC)为复杂性(拟度量)空间.

    注4  容易验证dCC上的一个拟度量.

    文献[2]证明了复杂性空间(CdC)是Smyth完备的.

    在复杂性理论中,算法的复杂性由它的运行时间函数f(n)(n${{\mathbb{N}}_{+}}$)来表示,其中f(n)被称为算法的复杂性函数.

    fgC,若对∀n${{\mathbb{N}}_{+}}$f(n)≤g(n),则称f的所有输入效率都优于g. 显然地,若f的所有输入效率都优于g,则dC(fg)=0.

    在本文中,设fgC,若存在n0${{\mathbb{N}}_{+}}$,使得对所有的nn0都有f(n)≤g(n),我们称f的渐近效率优于g.

    例2说明复杂性拟度量dC不适合刻画算法的渐近效率.

    例2  设fgC,对任意的n${{\mathbb{N}}_{+}}$f(n)=n+1,g(n)=$\frac{2^{n}}{n^{2}+2}$. 通过计算可得:当n=0,1,2,…,10时f(n)>g(n);当n≥11时f(n)<g(n). 因此,f的渐近效率优于g. 此时,显然dC(fg)≠0. 因此,复杂性拟度量dC不适合刻画算法的渐近效率.

2.   复杂性函数集上的模糊拟度量空间
  • 本节将在复杂性函数集C上引入模糊拟度量,以此刻画算法的渐近效率.

    定义8[9]  对∀fgCt>0,定义辅助函数QC

    其中t∈(nn+1],n${{\mathbb{N}}_{+}}$.

    性质1[9]  对∀fgCt∈(0,1],有QC(fgt)=dC(fg).

    性质2[9]  对∀fghCts>0,有QC(fgt+s)≤QC(fht)+QC(hgs).

    性质3[9]  对∀fgC,函数QC(fg,·)在(0,∞)上是非增的和左连续的.

    定理2  设C是复杂性函数集,*是连续t-模,fgC,在C×C×[0,∞)上定义函数MC

    (ⅰ) (MC,*)是C上的一个模糊拟度量;

    (ⅱ) 若(MdC,*)是由dC导出的标准模糊拟度量,则对∀t∈(0,1],MC(fgt)=MdC(fgt).

      (ⅰ) (FQM1)显然成立.

    (FQM2) 令fgC,对∀t>0,当f=g时,

    反之,若对∀t>0,

    则特别地当t=1时,

    因此

    又由性质1即得

    因为dCC上的拟度量,所以f=g.

    (FQM3) 令fghC,对∀ts≥0,不妨设MC(fht)≤MC(ghs),所以

    又因为

    即证得

    因此

    (FQM4) 由性质3可知,显然成立.

    (ⅱ) 由性质1,对∀fgC,当t∈(0,1]时, QC(fgt)=dC(fg),于是MC(fgt)=MdC(fgt).

    下文中的模糊拟度量空间(CMC,*)均为定理2中引入的模糊拟度量空间.

    定理3  在模糊拟度量空间(CMC,*)中,对∀fgCf的渐近效率优于g当且仅当存在n0${{\mathbb{N}}_{+}}$,使得对∀tn0,有MC(fgt)=1.

     必要性对∀fgC,若f的渐近效率优于g,则存在n0${{\mathbb{N}}_{+}}$,使得对∀nn0f(n)≤g(n). 由QC(fgt)的定义可得,对任意的tn0QC(fgt)=0,即MC(fgt)=1.

    充分性 由条件,当t∈(n0n0+1]时,MC(fgt)=1,即QC(fgt)=0. 于是,对∀nn0,有f(n)≤g(n),即f的渐近效率优于g.

    例3表明模糊拟度量MC可以刻画算法的渐近效率.

    例3  令f(n),g(n)为例2中的函数. 经计算得:当n=0,1,2,…,10时f(n)>g(n);当n≥11时f(n)<g(n). 因此,取n0=11,则对∀t∈(11,12],QC(fgt)=0. 由性质3可知函数QC(fg,·)是非增的,所以对∀t>11,QC(fgt)=0,即MC(fgt)=1. 因此f的渐近效率优于g. 即模糊拟度量MC可以刻画算法的渐近效率.

    定理4  模糊拟度量空间(CMC,*)是Smyth完备的.

     令{fn}是(CMC,*)中的左K-柯西列,取$\varepsilon \in\left(0, \frac{1}{2}\right)$,则存在n0${{\mathbb{N}}_{+}}$,使得当mnn0时,MC(fnfmt)>1-ε. 由定理2得MC(fnfm,1)=MdC(fnfm,1). 所以当mnn0时,

    因此,{fn}是(CdC)中的左K-柯西列. 因为(CdC)是Smyth完备的,所以存在fC,使得

    又由性质1可得,对∀t>0都有

    所以左K-柯西列{fn}是收敛的. 因此,(CMC,*)是Smyth完备的.

3.   不动点定理及其应用
  • 本节主要介绍模糊拟度量空间(CMC,*)的不动点定理及其应用.

    定义9  设Φ是模糊拟度量空间(CMC,*)上的一个自映射,对于fgCt∈(0,1],若存在k∈(0,1),使得

    则称自映射Φ是(0,1]-压缩的.

    定理5  若Φ是一个(0,1]-压缩的自映射,则Φ有唯一的不动点.

      令Φfn=fn+1,则存在k∈(0,1),使得对∀n${{\mathbb{N}}_{+}}$t∈(0,1],总有

    由性质1可得

    根据三角不等式得,对∀nm${{\mathbb{N}}_{+}}$

    所以{fn}在(CdC)中是一个左K-柯西列. 因为(CdC)是Smyth完备的,从而{fn}在度量(dC)s意义下是收敛的. 因此,序列{fn}在模糊度量(MdC)i意义下是收敛的. 即存在pC,对∀t∈(0,1],有

    又由定理2可得,对∀t∈(0,1],

    因此,{fn}在模糊度量MCi意义下收敛到p. 由Φ是(0,1]-压缩映射可得,对∀t∈(0,1],

    因此,{fn}在模糊度量MCi意义下收敛到Φp. 又因为模糊度量MCi的拓扑是T2分离的,所以p=Φp,即Φ有不动点p.

    下面证Φ的不动点的唯一性. 令qCΦ的不动点,则对∀t∈(0,1],MC(pqt)=MC(ΦpΦqt). 由于Φ是(0,1]-压缩的,因此存在k∈(0,1),使得对∀t∈(0,1],

    MC(pqt)=1. 因为MC(pq,·)是递增的,所以对∀t>0,都有MC(pqt)=1. 同理可得,对∀t>0,都有MC(qpt)=1. 因此p=q. 即Φ的不动点是唯一的.

    接下来我们将应用定理5来证明与分治算法相关的递归方程的解的存在唯一性.

    正如文献[1, 14]中所述,分治算法都是通过递归的方法将原始问题分解成几个子问题来解决,每个子问题都是由相同的算法单独解决的,然后组合后的结果就是原始问题的答案. 因此,分治算法的复杂性通常由下面的递归方程的解来表示:

    其中,n∈{bpp${{\mathbb{N}}_{+}}$},并且h(n)<+∞. 递归方程(1)自然地诱导出一个映射Φ,定义如下:

    文献[1]证明了:对∀fgC,当k=$\frac{1}{a}$时,dC(ΦfΦg)≤kdC(fg). 由性质1可得,对∀t∈(0,1],

    因此,Φ是一个(0,1]-压缩的自映射. 应用定理5可得,Φ有唯一的不动点g0C,即为递归方程(1)的解.

    最后我们将应用定理5来证明与快速排序算法相关的递归方程解的存在唯一性.

    文献[15]在讨论快速排序算法的平均案例分析时得出了一个递归方程T

    正如文献[16]中所述,递归方程(2)可以自然地诱导出一个映射ΦCC,定义如下:

    由文献[17]可知,对∀fgCdC(ΦfΦg)≤$\frac{d_{C}(f, g)}{2}$. 取k=$\frac{1}{2}$,由性质1得,对∀t∈(0,1],

    因此,Φ是一个(0,1]-压缩的自映射. 应用定理5可得,Φ有唯一的不动点g0C. 从而,若定义函数h${{\mathbb{N}}_{+}}$→[0,+∞)如下:

    h就是递归方程(2)的唯一解.

Reference (17)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return