Basic study on efficiency of the phenomenon reproduction type of Monte Carlo calculation and its application to the problems of disaster prevention

2018 Fiscal Year Final Research Report

• Project/Area Number: 15K04767
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Computational science
• Research Institution: Anan National College of Technology
• Principal Investigator: MATSUHO Shigeyuki 阿南工業高等専門学校, 創造技術工学科, 教授 (90157347)
• Co-Investigator(Kenkyū-buntansha): 西村 伸一 岡山大学, 環境生命科学研究科, 教授 (30198501)
• Research Collaborator: HIRAYAMA Motoi
• Project Period (FY): 2015-04-01 – 2019-03-31
• Keywords: モンテカルロ法 / シミュレーション / 擬似乱数 / 効率化計算

Outline of Final Research Achievements

This study developed precise pseudo-random numbers (p.r.n.) to perform efficient Monte Carlo (MC) simulation. The MC simulation is the powerful last means to solve a difficult problem. However, the necessary solution may not be provided if its calculation is inefficient. In addition, the computational result of the bad precision may lead a wrong engineering judgment. So, efficient computation is important. The efficient MC simulation is feasible by using good p.r.n. of the precision. In the proposed method, an error between the frequency distribution of the generated p.r.n. and its required probabilistic distribution is considered. The errors of many series of p.r.n. are observed and the p.r.n. seed minimizing the error is compiled into a database. In the MC calculation, a p.r.n. series of the seed registered with the database is used. This study formulated the error based on probabilistic moments. The effectiveness of the developed technique was illustrated by numerical examples.

Free Research Field

土木工学、構造工学、構造信頼性工学

Academic Significance and Societal Importance of the Research Achievements

モンテカルロ・シミュレーションは人命に関わる問題・膨大な予算や財産に関わる問題・問題解決に長期間を必要とする問題などの難解問題を扱う手法として極めて有力な最終手段である。本研究は、高精度計算の実現を可能とすることによって、実用的時間内に実用的な精度で解決策を得ることができるモンテカルロ・シミュレーションの開発に関する研究である。