Quasi-Monte Carlo simulation for the design of high-dimensional systems

2016 Fiscal Year Final Research Report

• Project/Area Number: 25350439
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Social systems engineering/Safety system
• Research Institution: National Graduate Institute for Policy Studies
• Principal Investigator: Morohosi Hozumi 政策研究大学院大学, 政策研究科, 教授 (10272387)
• Co-Investigator(Renkei-kenkyūsha): KOMAKI Fumiyasu 東京大学, 情報理工学系研究科, 教授 (70242039)
• Research Collaborator: FURUTA Takehiro
• Project Period (FY): 2013-04-01 – 2017-03-31
• Keywords: シミュレーション / OR / システム分析

Outline of Final Research Achievements

This research aims at advancing the use of quasi-random numbers for systems analysis and design, providing new algorithms and testing them by several numerical experiments. Problems studied in the research are, a particle filter using quasi-random numbers, an optimization algorithm for parameter identification, and random sample generation which follows very complicated high-dimensional probability distribution. In all the cases, we design algorithms using quasi-random numbers carefully so as to keep the good property of them, i.e. very precise uniformity in the region. We check the availability and superiority of quasi-random numbers for several numerical experiments, and find replacing random numbers by quasi-random ones often improves the computational performance of algorithms.

Free Research Field

社会システム工学