• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

A Development of Perfect Sampling with SAT/SMT Solvers

Research Project

Project/Area Number 20K11694
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Review Section Basic Section 60020:Mathematical informatics-related
Research InstitutionHiroshima University

Principal Investigator

Okamura Hiroyuki  広島大学, 先進理工系科学研究科(工), 教授 (10311812)

Project Period (FY) 2020-04-01 – 2023-03-31
Project Status Completed (Fiscal Year 2022)
Budget Amount *help
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2022: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2021: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2020: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
Keywordsシミュレーション / パーフェクトサンプリング / マルコフ過程 / SAT/SMTソルバ
Outline of Research at the Start

パーフェクトサンプリングとはマルコフ連鎖の定常分布に従うサンプルを厳密にサンプリングする手法であり,数理的には無限時間のシミュレーションを有限時間で実行することができる.SAT/SMTソルバを用いたパーフェクトサンプリング手法は申請者が考案した手法であり,パーフェクトサンプリングの適用範囲を格段に向上させた.本研究では,SAT/SMTソルバによるパーフェクトサンプリング手法の適用範囲をさらに拡大するために,一般化確率ペトリネット(GSPN),マルコフ再生確率ペトリネット(MRSPN),流体確率ペトリネット(FSPN)に対する手法を開発する.

Outline of Final Research Achievements

In this research, we developed a highly versatile perfect sampling algorithm using SAT/SMT solvers. Perfect sampling is a method drawing samples from the stationary distribution of a Markov chain exactly. The coupling from the Past (CFTP) method is known as one of the representative method of perfect sampling. Perfect sampling is mathematically interesting because it provides a framework for simulating an infinite-time Markov chain in finite time. In this research, we proposed an algorithm to automatically extract formulas to be solved by SAT/SMT solvers for Markov chain models described by stochastic Petri nets or generalized stochastic Petri nets, and generate samples that exactly follow the stationary distribution. Furthermore, we also proposed a method to enhance the computational time, which is a weakness of conventional perfect sampling algorithms, and achieved a speedup of nearly 10,000 times in practice.

Academic Significance and Societal Importance of the Research Achievements

パーフェクトサンプリングアルゴリズムは無限時間を有限時間で厳密にシミュレートするための手法であり,その応用範囲はシミュレーションによるシステム性能評価だけでなく,機械学習に対する学習アルゴリズム等への利用など多岐にわたる.これまでのアルゴリズムは実用面については特殊な構造を持つ確率過程にしか適用できなかったが,本研究ではモデルの数理的な構造を自動的に抽出することで,かなり多くのモデルに対して汎用的に利用できるアルゴリズムを開発した.

Report

(4 results)
  • 2022 Annual Research Report   Final Research Report ( PDF )
  • 2021 Research-status Report
  • 2020 Research-status Report
  • Research Products

    (10 results)

All 2022 2021 2020

All Journal Article (7 results) (of which Int'l Joint Research: 1 results,  Peer Reviewed: 5 results) Presentation (3 results)

  • [Journal Article] Parameter Estimation of Markovian Arrivals with Utilization Data2022

    • Author(s)
      Chen Li, Junjun Zheng, Hiroyuki Okamura, Tadashi Dohi
    • Journal Title

      IEICE Transactions on Communications

      Volume: E105.B Issue: 1 Pages: 1-10

    • DOI

      10.1587/transcom.2021EBP3007

    • NAID

      130008138811

    • ISSN
      0916-8516, 1745-1345
    • Year and Date
      2022-01-01
    • Related Report
      2022 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Hierarchical Bayesian Parameter Estimation of Queueing Systems using Utilization Data2022

    • Author(s)
      Chen Li, Junjun Zheng, Hiroyuki Okamura, Tadashi Dohi
    • Journal Title

      Int. J. Perform. Eng.

      Volume: 18 Issue: 5 Pages: 307-316

    • DOI

      10.23940/ijpe.22.05.p1.307316

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Age replacement with Markovian opportunity process2021

    • Author(s)
      Zheng Junjun、Okamura Hiroyuki、Dohi Tadashi
    • Journal Title

      Reliability Engineering & System Safety

      Volume: 216 Pages: 107949-107949

    • DOI

      10.1016/j.ress.2021.107949

    • Related Report
      2021 Research-status Report
  • [Journal Article] Quantitative Security Evaluation of Intrusion Tolerant Systems With Markovian Arrivals2021

    • Author(s)
      Zheng Junjun、Okamura Hiroyuki、Dohi Tadashi、Trivedi Kishor S.
    • Journal Title

      IEEE Transactions on Reliability

      Volume: 70 Issue: 2 Pages: 547-562

    • DOI

      10.1109/tr.2020.3026570

    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Int'l Joint Research
  • [Journal Article] Availability importance measures of components in smart electric power grid systems2021

    • Author(s)
      Zheng Junjun、Okamura Hiroyuki、Pang Taoming、Dohi Tadashi
    • Journal Title

      Reliability Engineering & System Safety

      Volume: 205 Pages: 107164-107164

    • DOI

      10.1016/j.ress.2020.107164

    • Related Report
      2021 Research-status Report
  • [Journal Article] A phase expansion for non-Markovian availability models with time-based aperiodic rejuvenation and checkpointing2020

    • Author(s)
      Junjun Zheng, Hiroyuki Okamura, Tadashi Dohi
    • Journal Title

      Communications in Statistics - Theory and Methods

      Volume: 49 Issue: 15 Pages: 3712-3729

    • DOI

      10.1080/03610926.2019.1708400

    • Related Report
      2020 Research-status Report
    • Peer Reviewed
  • [Journal Article] An SMT-Based Perfect Sampling Algorithm for Stochastic Petri Nets2020

    • Author(s)
      Hiroyuki Okamura, Kazuya Morihara Tadashi Dohi
    • Journal Title

      Proceedings of the 13th EAI International Conference on Performance Evaluation Methodologies and Tools

      Volume: - Pages: 104-111

    • DOI

      10.1145/3388831.3388844

    • Related Report
      2020 Research-status Report
    • Peer Reviewed
  • [Presentation] パーフェクトサンプリングアルゴリズムの確率的モデルチェッキングへの応用に関する一考察2022

    • Author(s)
      安藤清佳,岡村寛之,土肥正
    • Organizer
      日本オペレーションズリサーチ学会中国・四国支部SSOR
    • Related Report
      2022 Annual Research Report
  • [Presentation] パーフェクトサンプリングアルゴリズムの高速化に関する考察2022

    • Author(s)
      安藤清佳,岡村寛之,土肥正
    • Organizer
      日本オペレーションズリサーチ学会「信頼性とOR」研究部会
    • Related Report
      2022 Annual Research Report
  • [Presentation] On perfect sampling algorithm for generalized stochastic Petri nets2022

    • Author(s)
      Hiroyuki Okamura, Tadashi Dohi
    • Organizer
      2021年度待ち行列シンポジウム
    • Related Report
      2021 Research-status Report

URL: 

Published: 2020-04-28   Modified: 2024-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi