Theoretical Computer Science最新文献

{"title":"Asynchronous fully-decentralized SGD in the cluster-based model","authors":"Hagit Attiya ,&nbsp;Noa Schiller","doi":"10.1016/j.tcs.2025.115073","DOIUrl":"10.1016/j.tcs.2025.115073","url":null,"abstract":"<div><div>This paper presents fault-tolerant asynchronous <em>Stochastic Gradient Descent</em> (<em>SGD</em>) algorithms. SGD is widely used for approximating the minimum of a cost function <em>Q</em>, a core part of optimization and learning algorithms. Our algorithms are designed for the <em>cluster-based</em> model, which combines message-passing and shared-memory communication layers. Processes may fail by <em>crashing</em>, and the algorithm inside each cluster is <em>wait-free</em>, using only reads and writes.</div><div>For a <em>strongly convex Q</em>, our algorithm <em>can withstand partitions of the system</em>. It provides convergence rate that is the maximal distributed acceleration over the optimal convergence rate of <em>sequential</em> SGD.</div><div>For arbitrary smooth functions, the convergence rate has an additional term that depends on the maximal difference between the parameters at the same iteration. (This holds under standard assumptions on <em>Q</em>.) In this case, the algorithm obtains the same convergence rate as sequential SGD, up to a logarithmic factor. This is achieved by using, at each iteration, a <em>multidimensional approximate agreement</em> algorithm, tailored for the cluster-based model.</div><div>The general algorithm communicates with nonfaulty processes belonging to clusters that include a majority of all processes. We prove that this condition is necessary when optimizing some non-convex functions.</div></div>","PeriodicalId":49438,"journal":{"name":"Theoretical Computer Science","volume":"1031 ","pages":"Article 115073"},"PeriodicalIF":0.9,"publicationDate":"2025-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143265843","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The (conditional) matroidal connectivity of varietal hypercube","authors":"Lulu Yang ,&nbsp;Shuming Zhou ,&nbsp;Qifan Zhang ,&nbsp;Guanqin Lian","doi":"10.1016/j.tcs.2025.115076","DOIUrl":"10.1016/j.tcs.2025.115076","url":null,"abstract":"<div><div>As multiprocessor systems continue to grow in scale, their underlying interconnection networks face increasingly challenging issues to characterize reliability and fault tolerance. Matroidal connectivity and conditional matroidal connectivity are two novel metrics that enhance network reliability by dividing the edge set into several parts and flexibly adjusting the number of faulty edges in each dimension to meet practical requirements. The <em>n</em>-dimensional varietal hypercube <span><math><mi>V</mi><msub><mrow><mi>Q</mi></mrow><mrow><mi>n</mi></mrow></msub></math></span> is vertex-transitive and edge-transitive, and it is a variant of the traditional hypercube. In this paper, we first determine the matroidal connectivity and conditional matroidal connectivity of <span><math><mi>V</mi><msub><mrow><mi>Q</mi></mrow><mrow><mi>n</mi></mrow></msub></math></span>. Then we conduct a comparative analysis of <span><math><mi>V</mi><msub><mrow><mi>Q</mi></mrow><mrow><mi>n</mi></mrow></msub></math></span>'s matroidal connectivity with some kinds of conditional connectivities, namely <em>g</em>-extra edge connectivity, <em>g</em>-component edge connectivity, and <em>g</em>-super edge connectivity. Our results show that matroidal connectivity outperforms the conditional edge connectivities above, underscoring its efficiency in bolstering the fault tolerance of interconnection networks under specific conditions.</div></div>","PeriodicalId":49438,"journal":{"name":"Theoretical Computer Science","volume":"1030 ","pages":"Article 115076"},"PeriodicalIF":0.9,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143305540","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Oracle separations for non-adaptive collapse-free quantum computing","authors":"Henrique Hepp ,&nbsp;Murilo V.G. da Silva ,&nbsp;Leandro M. Zatesko","doi":"10.1016/j.tcs.2025.115078","DOIUrl":"10.1016/j.tcs.2025.115078","url":null,"abstract":"<div><div>Aaronson et al. (2016) introduced the quantum complexity class <span><math><mi>naCQP</mi></math></span> (<em>non-adaptive Collapse-free Quantum Polynomial time</em>), also known as <span><math><mi>PDQP</mi></math></span> (<em>Product Dynamical Quantum Polynomial time</em>), aiming to define a class larger than <span><math><mi>BQP</mi></math></span> (<em>Bounded-error Quantum Polynomial time</em>), but not large enough to include <span><math><mi>NP</mi></math></span>-complete problems. Aaronson et al. showed that <span><math><mi>SZK</mi></math></span> (<em>Statistical Zero Knowledge</em>) is contained in <span><math><mi>naCQP</mi></math></span> and that there is an oracle <em>A</em> for which <span><math><msup><mrow><mi>NP</mi></mrow><mrow><mi>A</mi></mrow></msup><mo>⊈</mo><msup><mrow><mi>naCQP</mi></mrow><mrow><mi>A</mi></mrow></msup></math></span>. We prove that: there is an oracle <em>A</em> for which <span><math><msup><mrow><mi>P</mi></mrow><mrow><mi>A</mi></mrow></msup><mo>=</mo><msup><mrow><mi>BQP</mi></mrow><mrow><mi>A</mi></mrow></msup><mo>=</mo><msup><mrow><mi>SZK</mi></mrow><mrow><mi>A</mi></mrow></msup><mo>=</mo><msup><mrow><mi>naCQP</mi></mrow><mrow><mi>A</mi></mrow></msup><mo>≠</mo><msup><mrow><mo>(</mo><mi>UP</mi><mo>∩</mo><mi>coUP</mi><mo>)</mo></mrow><mrow><mi>A</mi></mrow></msup><mo>=</mo><msup><mrow><mi>EXP</mi></mrow><mrow><mi>A</mi></mrow></msup></math></span>, where <span><math><mi>UP</mi></math></span> (<em>Unambiguous Polynomial time</em>) is an important subclass of <span><math><mi>NP</mi></math></span>; relative to an oracle <em>A</em> chosen uniformly at random, it holds <span><math><msup><mrow><mo>(</mo><mi>UP</mi><mo>∩</mo><mi>coUP</mi><mo>)</mo></mrow><mrow><mi>A</mi></mrow></msup><mo>⊈</mo><msup><mrow><mi>naCQP</mi></mrow><mrow><mi>A</mi></mrow></msup></math></span> with probability 1. Our results are a strengthening of the results by Bennett et al. (1997), Fortnow and Rogers (1999), Tamon and Yamakami (2001), and Aaronson et al. (2016).</div></div>","PeriodicalId":49438,"journal":{"name":"Theoretical Computer Science","volume":"1030 ","pages":"Article 115078"},"PeriodicalIF":0.9,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143305538","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"On the optimal mixing problem of approximate Nash equilibria in bimatrix games","authors":"Xiaotie Deng ,&nbsp;Dongchen Li ,&nbsp;Hanyu Li","doi":"10.1016/j.tcs.2025.115072","DOIUrl":"10.1016/j.tcs.2025.115072","url":null,"abstract":"<div><div>This paper introduces the optimal mixing problem, a natural extension of the computation of approximate Nash Equilibria (NE) in bimatrix games. The problem focuses on determining the optimal convex combination of given strategies that minimizes the approximation (i.e., regret) in NE computation. We develop algorithms for the exact and approximate optimal mixing problems and present new complexity results that bridge both practical and theoretical aspects of NE computation. Practically, our algorithms can be used to enhance and integrate arbitrary existing constant-approximate NE algorithms, offering a powerful tool for the design of approximate NE algorithms. Theoretically, these algorithms allow us to explore the implications of support restrictions on approximate NE and derive the upper-bound separations between approximate NE and exact NE. Consequently, this work contributes to theoretical understandings of the computational complexity of approximate NE under various constraints and practical improvements in multi-agent reinforcement learning (MARL) and other fields where NE computation is involved.</div></div>","PeriodicalId":49438,"journal":{"name":"Theoretical Computer Science","volume":"1031 ","pages":"Article 115072"},"PeriodicalIF":0.9,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143265810","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Maximum independent set formation on a finite grid by myopic robots","authors":"Raja Das,&nbsp;Avisek Sharma,&nbsp;Buddhadeb Sau","doi":"10.1016/j.tcs.2025.115077","DOIUrl":"10.1016/j.tcs.2025.115077","url":null,"abstract":"<div><div>This work deals with the Maximum Independent Set (<span><math><mi>MAXIS</mi></math></span>) formation problem in a finite rectangular grid by autonomous robots. Suppose we are given a set of identical robots, where each robot is placed on a node of a finite rectangular grid <span><math><mi>G</mi></math></span> such that no two robots are on the same node. The <span><math><mi>MAXIS</mi></math></span> formation problem asks to design an algorithm and each robot will move autonomously after executing the algorithm and terminate at a node such that after a finite time the set of nodes occupied by the robots is a maximum independent set of <span><math><mi>G</mi></math></span>. We assume that robots are anonymous and they execute the same distributed algorithm.</div><div>Previous works solved this problem using one or several door nodes through which the robots enter the grid or the graph one by one and occupy the required nodes. In this work, we propose a deterministic algorithm that solves the <span><math><mi>MAXIS</mi></math></span> formation problem in a more generalized scenario, i.e., when the total number of required robots to form a <span><math><mi>MAXIS</mi></math></span> are arbitrarily placed on the grid. The proposed algorithm works under a semi-synchronous scheduler using robots with only two hop visibility range and only three colors.</div></div>","PeriodicalId":49438,"journal":{"name":"Theoretical Computer Science","volume":"1031 ","pages":"Article 115077"},"PeriodicalIF":0.9,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143265844","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Private least absolute deviations with heavy-tailed data","authors":"Di Wang ,&nbsp;Jinhui Xu","doi":"10.1016/j.tcs.2025.115071","DOIUrl":"10.1016/j.tcs.2025.115071","url":null,"abstract":"<div><div>We study the problem of Differentially Private Stochastic Convex Optimization (DPSCO) with heavy-tailed data. Specifically, we focus on the problem of Least Absolute Deviations, i.e., <span><math><msub><mrow><mi>ℓ</mi></mrow><mrow><mn>1</mn></mrow></msub></math></span>-norm linear regression, in the <em>ϵ</em>-DP model. While most previous work focuses on the case where the loss function is Lipschitz, in this paper we only need to assume the variates have bounded moments. Firstly, we study the case where the <span><math><msub><mrow><mi>ℓ</mi></mrow><mrow><mn>2</mn></mrow></msub></math></span> norm of data has a bounded second-order moment. We propose an algorithm that is based on the exponential mechanism and show that it is possible to achieve an upper bound of <span><math><mover><mrow><mi>O</mi></mrow><mrow><mo>˜</mo></mrow></mover><mo>(</mo><msqrt><mrow><mfrac><mrow><mi>d</mi></mrow><mrow><mi>n</mi><mi>ϵ</mi></mrow></mfrac></mrow></msqrt><mo>)</mo></math></span> (with high probability). Next, we relax the assumption to bounded <em>θ</em>-th order moment with some <span><math><mi>θ</mi><mo>∈</mo><mo>(</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>)</mo></math></span> and show that it is possible to achieve an upper bound of <span><math><mover><mrow><mi>O</mi></mrow><mrow><mo>˜</mo></mrow></mover><mo>(</mo><msup><mrow><mo>(</mo><mfrac><mrow><mi>d</mi></mrow><mrow><mi>n</mi><mi>ϵ</mi></mrow></mfrac><mo>)</mo></mrow><mrow><mfrac><mrow><mi>θ</mi><mo>−</mo><mn>1</mn></mrow><mrow><mi>θ</mi></mrow></mfrac></mrow></msup><mo>)</mo></math></span>. Our algorithms can also be extended to more relaxed cases where only each coordinate of the data has bounded moments, and we can get an upper bound of <span><math><mover><mrow><mi>O</mi></mrow><mrow><mo>˜</mo></mrow></mover><mo>(</mo><mfrac><mrow><mi>d</mi></mrow><mrow><msqrt><mrow><mi>n</mi><mi>ϵ</mi></mrow></msqrt></mrow></mfrac><mo>)</mo></math></span> and <span><math><mover><mrow><mi>O</mi></mrow><mrow><mo>˜</mo></mrow></mover><mo>(</mo><mfrac><mrow><mi>d</mi></mrow><mrow><msup><mrow><mo>(</mo><mi>n</mi><mi>ϵ</mi><mo>)</mo></mrow><mrow><mfrac><mrow><mi>θ</mi><mo>−</mo><mn>1</mn></mrow><mrow><mi>θ</mi></mrow></mfrac></mrow></msup></mrow></mfrac><mo>)</mo></math></span> in the second and <em>θ</em>-th moment case respectively.</div></div>","PeriodicalId":49438,"journal":{"name":"Theoretical Computer Science","volume":"1030 ","pages":"Article 115071"},"PeriodicalIF":0.9,"publicationDate":"2025-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143305539","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"On the computational and descriptional complexity of multi-pattern languages","authors":"Jingnan Xie ,&nbsp;Harry B. Hunt III ,&nbsp;Richard E. Stearns","doi":"10.1016/j.tcs.2025.115063","DOIUrl":"10.1016/j.tcs.2025.115063","url":null,"abstract":"<div><div>We study the complexity of a restricted version of the universality problem: testing equivalence to <span><math><msup><mrow><mo>{</mo><mn>0</mn><mo>,</mo><mn>1</mn><mo>}</mo></mrow><mrow><mo>⁎</mo></mrow></msup></math></span> for languages that are either <span><math><msup><mrow><mo>{</mo><mn>0</mn><mo>,</mo><mn>1</mn><mo>}</mo></mrow><mrow><mo>⁎</mo></mrow></msup></math></span> or <span><math><msup><mrow><mo>{</mo><mn>0</mn><mo>,</mo><mn>1</mn><mo>}</mo></mrow><mrow><mo>⁎</mo></mrow></msup><mo>−</mo><mo>{</mo><mi>w</mi><mo>}</mo></math></span> where <span><math><mi>w</mi><mo>∈</mo><msup><mrow><mo>{</mo><mn>0</mn><mo>,</mo><mn>1</mn><mo>}</mo></mrow><mrow><mo>⁎</mo></mrow></msup></math></span>. We show that this restricted problem for multi-pattern languages (MPL) is NP-hard, and for multi-pattern languages with regular substitutions (MPL_REG) is not recursively enumerable. Moreover, this undecidability result holds for any class of languages that effectively contains <span><math><mo>{</mo><mi>w</mi><mi>#</mi><mi>w</mi><mspace></mspace><mo>|</mo><mspace></mspace><mi>w</mi><mo>∈</mo><msup><mrow><mo>{</mo><mn>0</mn><mo>,</mo><mn>1</mn><mo>}</mo></mrow><mrow><mo>⁎</mo></mrow></msup><mo>}</mo></math></span>, and is effectively closed under union, and concatenation with regular sets. Consequently, it has broad applicability and extends to many generalizations of regular languages. Sufficient conditions are then presented for a language predicate to be as hard as the restricted universality problem. By applying these conditions, we develop a uniform method for showing undecidability and complexity results simultaneously via highly efficient many-one reductions. In addition, the properties of this restricted universality problem can be used to investigate the descriptional complexity of multi-patterns. We show that the trade-off between multi-patterns and regular expressions is not exponential-size bounded. Non-recursive trade-offs between multi-patterns (or multi-patterns with regular substitutions) and numerous classes of language descriptors are also established.</div></div>","PeriodicalId":49438,"journal":{"name":"Theoretical Computer Science","volume":"1030 ","pages":"Article 115063"},"PeriodicalIF":0.9,"publicationDate":"2025-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143306088","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluating the reliability of complete Josephus cubes under extra link fault with the optimal solution of the edge isoperimetric problem","authors":"Sufang Liu ,&nbsp;Zhaoman Huang ,&nbsp;Yueke Lv ,&nbsp;Chia-Wei Lee","doi":"10.1016/j.tcs.2025.115064","DOIUrl":"10.1016/j.tcs.2025.115064","url":null,"abstract":"<div><div>The <em>h</em>-extra edge-connectivity of a connected graph <em>G</em> <span><math><msub><mrow><mi>λ</mi></mrow><mrow><mi>h</mi></mrow></msub><mo>(</mo><mi>G</mi><mo>)</mo></math></span>, as a generalization of the classic Menger's Theorem, is the minimum number of edges that need to be removed to disconnect graph <em>G</em> and ensure that each component of the remaining graph has at least <em>h</em> vertices. This paper focuses on the reliability of the <em>h</em>-extra edge-connectivity of the complete Josephus cube <span><math><mi>C</mi><mi>J</mi><msub><mrow><mi>C</mi></mrow><mrow><mi>n</mi></mrow></msub></math></span> interconnection network, a variant of <span><math><msub><mrow><mi>Q</mi></mrow><mrow><mi>n</mi></mrow></msub></math></span>, as the underlying topology, denoted as <span><math><msub><mrow><mi>λ</mi></mrow><mrow><mi>h</mi></mrow></msub><mo>(</mo><mi>C</mi><mi>J</mi><msub><mrow><mi>C</mi></mrow><mrow><mi>n</mi></mrow></msub><mo>)</mo></math></span>. By analyzing the properties of the optimal solution of the edge isoperimetric problem on <span><math><mi>C</mi><mi>J</mi><msub><mrow><mi>C</mi></mrow><mrow><mi>n</mi></mrow></msub></math></span>, the exact value and the <span><math><msub><mrow><mi>λ</mi></mrow><mrow><mi>h</mi></mrow></msub></math></span>-optimality of the <span><math><msub><mrow><mi>λ</mi></mrow><mrow><mi>h</mi></mrow></msub><mo>(</mo><mi>C</mi><mi>J</mi><msub><mrow><mi>C</mi></mrow><mrow><mi>n</mi></mrow></msub><mo>)</mo></math></span> for <span><math><mn>1</mn><mo>≤</mo><mi>h</mi><mo>≤</mo><msup><mrow><mn>2</mn></mrow><mrow><mo>⌊</mo><mi>n</mi><mo>/</mo><mn>2</mn><mo>⌋</mo><mo>+</mo><mn>1</mn></mrow></msup></math></span> is characterized. For a sufficiently large positive integer <em>n</em>, about 44.44% of positive integers <em>h</em> in the well-defined interval <span><math><mn>1</mn><mo>≤</mo><mi>h</mi><mo>≤</mo><msup><mrow><mn>2</mn></mrow><mrow><mi>n</mi><mo>−</mo><mn>1</mn></mrow></msup></math></span>, the corresponding <span><math><msub><mrow><mi>λ</mi></mrow><mrow><mi>h</mi></mrow></msub><mo>(</mo><mi>C</mi><mi>J</mi><msub><mrow><mi>C</mi></mrow><mrow><mi>n</mi></mrow></msub><mo>)</mo></math></span> occurs a concentration phenomenon.</div></div>","PeriodicalId":49438,"journal":{"name":"Theoretical Computer Science","volume":"1030 ","pages":"Article 115064"},"PeriodicalIF":0.9,"publicationDate":"2025-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143306089","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Feasibility analysis of recurrent DAG tasks is PSPACE-hard","authors":"Vincenzo Bonifaci ,&nbsp;Alberto Marchetti-Spaccamela","doi":"10.1016/j.tcs.2024.115062","DOIUrl":"10.1016/j.tcs.2024.115062","url":null,"abstract":"<div><div>We study a popular task model for scheduling parallel real-time tasks, where the internal parallelism of each task is modeled by a directed acyclic graph (DAG). We show that deciding the feasibility of a set of sporadically recurrent DAG tasks is hard for the complexity class <span>PSPACE</span>, thus ruling out approaches to this problem that rely on Integer Linear Programming or Satisfiability solvers (assuming <span><math><mtext>NP</mtext><mo>≠</mo><mtext>PSPACE</mtext></math></span>).</div></div>","PeriodicalId":49438,"journal":{"name":"Theoretical Computer Science","volume":"1030 ","pages":"Article 115062"},"PeriodicalIF":0.9,"publicationDate":"2025-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143306090","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Large deviation properties for pattern statistics in primitive rational models","authors":"Massimiliano Goldwurm,&nbsp;Marco Vignati","doi":"10.1016/j.tcs.2024.115055","DOIUrl":"10.1016/j.tcs.2024.115055","url":null,"abstract":"<div><div>We present a large deviation property for pattern statistics representing the number of occurrences of a symbol in words of given length generated at random according to a rational stochastic model. This result is proved assuming that the transition matrix of the model is primitive. We show how the rate function of the large deviation property depends on the main eigenvalues and eigenvectors of the transition matrices associated with the different symbols of the alphabet. We also yield general conditions to guarantee that the range of validity of the large deviation estimate coincides with the whole interval <span><math><mo>(</mo><mn>0</mn><mo>,</mo><mn>1</mn><mo>)</mo></math></span>, which represents in our context the largest possible open interval where the property may hold. The case of smaller intervals of validity is finally examined by means of examples.</div></div>","PeriodicalId":49438,"journal":{"name":"Theoretical Computer Science","volume":"1030 ","pages":"Article 115055"},"PeriodicalIF":0.9,"publicationDate":"2025-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143306092","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}