WSQN2901 Agent-based Transport Simulation

広島大学シラバス

English

年度 2024年度開講部局先進理工系科学研究科博士課程前期先進理工系科学専攻理工学融合プログラム

講義コード WSQN2901 科目区分専門的教育科目

授業科目名 Agent-based Transport Simulation

授業科目名
（フリガナ）

英文授業科目名 Agent-based Transport Simulation

担当教員名 FOURIE PIETER JACOBUS

担当教員名
(フリガナ) フーリー　ピーター　ヤクバス

開講キャンパス東広島開設期 1年次生前期集中

曜日・時限・講義室 (集) 集中：メディアセンター本館2F端末室

授業の方法講義授業の方法
【詳細情報】　

講義中心、演習中心、板書多用、ディスカッション、学生の発表、野外実習、作業、薬品使用

単位 2.0 週時間　使用言語 E : 英語

学習の段階 7 : 大学院発展的レベル

学問分野（分野） 25 : 理工学

学問分野（分科） 13 : 土木工学

対象学生 Master / PhD students

授業のキーワード

教職専門科目　教科専門科目　

プログラムの中での
この授業科目の位置づけ
（学部生対象科目のみ）
到達度評価
の評価項目
（学部生対象科目のみ）

授業の目標・概要等 The class objectives of the Agent-based transport simulation course are to introduce students to the philosophy and theory behind modeling social systems, specifically in the context of transportation systems. The course aims to provide students with a comprehensive understanding of agent-based modeling approaches, as well as practical experience in using software tools such as NetLogo, Intellij IDEA, and MATSim to build and analyze transportation models.
Throughout the course, students will learn about a range of topics, including transportation networks and land use, population synthesis, primary activity location choice, activity scheduling, big data-driven approaches, and MATSim co-evolutionary engine. By the end of the course, students should be able to apply these concepts to develop and analyze their own agent-based transport simulations.

授業計画 Lecture 1,2
Theory:
Philosophy of modeling social systems, intro to agent-based modeling. Thomas Hobbes and assumptions about human nature. Emergence and path-dependence. The Law of Large Numbers. Scale-free phase transitions. Aggregate versus disaggregate models. Agent-based modelling approaches. Cellular automata. Object-oriented programming. Spatial models.
Practical:
Installation of NetLogo, toy model exploration. Experimentation and modification of NetLogo toy models.
Lecture 3
Theory:
Overview of agent-based transport simulation central topics. Four-step modelling vs activity-based travel demand modelling. Microsimulation vs. meso-scale.
Practical:
Installation of Intellij IDEA, MATSim, QGIS. Exploring and running sample MATSim data sets. Visualisation of agent-based simulation output.
Lecture 4
Theory:
Transportation networks and land use. Data sources. Accessibility. Space syntax, network topology. Modeling transit.
Practical:
Generating MATSim network, facilities and transit input files for a medium-size city, e.g. Sioux Falls.
Lecture 5,6
Theory:
Population synthesis: Data sources and formats. Imputing missing attributes. Generalized raking. Household construction. Land use and spatial distribution of households. Discrete choice modeling introduction. Modelling car ownership and licensure.
Practical:
Population synthesis in R and Java.
Lecture 7
Theory:
Primary activity location choice: data sources and formats. Predictors of activity location choice. Comparing random forest and discrete choice modeling of location choice.
Practical:
Modeling primary activity location choice in R and Java.
Lecture 8,9
Theory:
Activity scheduling. Overview of activity scheduling approaches, e.g. TASHA, ALBATROS. Secondary activity accessibility in activity scheduling. Space-time prisms and travel time budgets. Secondary activity location choice.
Practical:
Modeling activity scheduling in R and Java.
Lecture 10, 11
Theory:
Big data-driven approaches. Synthesizing individual travelers from big data aggregates. Generative graphical models. Remaining challenges to data-driven demand generation.
Practical:
Day traveler synthesis from aggregated big data histograms.
Lecture 12,13
Theory:
MATSim co-evolutionary engine. Replanning modules. Time mutation. Least-cost routing. Subtour mode choice. MATSim configuration. MATSim utility function. Transport mode and activity participation utility parameters.
Practical:
Running, visualizing and analyzing MATSim runs of previously generated travel demand.
Lecture 14, 15
Theory:
MATSim events. Simple events analysis code. Overview of important MATSim ‘contrib’ modules: autonomous vehicles, emissions, dynamically routed transit.
Practical:
Develop a customized replanning module for MATSim and an events handler to analyze its effects.

【Schedule】 Aug. 19-23 from 8:45-14:20

教科書・参考書等 Ball, Philip. Critical mass: How one thing leads to another. Macmillan, 2004.
Axhausen, Kay Werner, Andreas Horni, and Kai Nagel. The multi-agent transport simulation MATSim. Ubiquity Press, 2016.

授業で使用する
メディア・機器等　

【詳細情報】

授業で取り入れる
学習手法　

予習・復習への
アドバイス Please bring your own laptop in every lecture

履修上の注意
受講条件等

成績評価の基準等 Each topic covered will present a small coding challenge. The aggregate grade will be computed from these challenges, each weighted by the number of lectures consumed by the topic

実務経験

実務経験の概要と
それに基づく授業内容

メッセージ

その他

すべての授業科目において，授業改善アンケートを実施していますので，回答に協力してください。
回答に対しては教員からコメントを入力しており，今後の改善につなげていきます。

年度	2024年度	開講部局	先進理工系科学研究科博士課程前期先進理工系科学専攻理工学融合プログラム
講義コード	WSQN2901	科目区分	専門的教育科目
授業科目名	Agent-based Transport Simulation
授業科目名（フリガナ）
英文授業科目名	Agent-based Transport Simulation
担当教員名	FOURIE PIETER JACOBUS
担当教員名 (フリガナ)	フーリー　ピーター　ヤクバス
開講キャンパス	東広島	開設期	1年次生前期集中
曜日・時限・講義室	(集) 集中：メディアセンター本館2F端末室
授業の方法	講義	授業の方法【詳細情報】
講義中心、演習中心、板書多用、ディスカッション、学生の発表、野外実習、作業、薬品使用
単位	2.0	週時間		使用言語	E : 英語
学習の段階	7 : 大学院発展的レベル
学問分野（分野）	25 : 理工学
学問分野（分科）	13 : 土木工学
対象学生	Master / PhD students
授業のキーワード
教職専門科目		教科専門科目
プログラムの中でのこの授業科目の位置づけ（学部生対象科目のみ）
到達度評価の評価項目（学部生対象科目のみ）
授業の目標・概要等	The class objectives of the Agent-based transport simulation course are to introduce students to the philosophy and theory behind modeling social systems, specifically in the context of transportation systems. The course aims to provide students with a comprehensive understanding of agent-based modeling approaches, as well as practical experience in using software tools such as NetLogo, Intellij IDEA, and MATSim to build and analyze transportation models. Throughout the course, students will learn about a range of topics, including transportation networks and land use, population synthesis, primary activity location choice, activity scheduling, big data-driven approaches, and MATSim co-evolutionary engine. By the end of the course, students should be able to apply these concepts to develop and analyze their own agent-based transport simulations.
授業計画	Lecture 1,2 Theory: Philosophy of modeling social systems, intro to agent-based modeling. Thomas Hobbes and assumptions about human nature. Emergence and path-dependence. The Law of Large Numbers. Scale-free phase transitions. Aggregate versus disaggregate models. Agent-based modelling approaches. Cellular automata. Object-oriented programming. Spatial models. Practical: Installation of NetLogo, toy model exploration. Experimentation and modification of NetLogo toy models. Lecture 3 Theory: Overview of agent-based transport simulation central topics. Four-step modelling vs activity-based travel demand modelling. Microsimulation vs. meso-scale. Practical: Installation of Intellij IDEA, MATSim, QGIS. Exploring and running sample MATSim data sets. Visualisation of agent-based simulation output. Lecture 4 Theory: Transportation networks and land use. Data sources. Accessibility. Space syntax, network topology. Modeling transit. Practical: Generating MATSim network, facilities and transit input files for a medium-size city, e.g. Sioux Falls. Lecture 5,6 Theory: Population synthesis: Data sources and formats. Imputing missing attributes. Generalized raking. Household construction. Land use and spatial distribution of households. Discrete choice modeling introduction. Modelling car ownership and licensure. Practical: Population synthesis in R and Java. Lecture 7 Theory: Primary activity location choice: data sources and formats. Predictors of activity location choice. Comparing random forest and discrete choice modeling of location choice. Practical: Modeling primary activity location choice in R and Java. Lecture 8,9 Theory: Activity scheduling. Overview of activity scheduling approaches, e.g. TASHA, ALBATROS. Secondary activity accessibility in activity scheduling. Space-time prisms and travel time budgets. Secondary activity location choice. Practical: Modeling activity scheduling in R and Java. Lecture 10, 11 Theory: Big data-driven approaches. Synthesizing individual travelers from big data aggregates. Generative graphical models. Remaining challenges to data-driven demand generation. Practical: Day traveler synthesis from aggregated big data histograms. Lecture 12,13 Theory: MATSim co-evolutionary engine. Replanning modules. Time mutation. Least-cost routing. Subtour mode choice. MATSim configuration. MATSim utility function. Transport mode and activity participation utility parameters. Practical: Running, visualizing and analyzing MATSim runs of previously generated travel demand. Lecture 14, 15 Theory: MATSim events. Simple events analysis code. Overview of important MATSim ‘contrib’ modules: autonomous vehicles, emissions, dynamically routed transit. Practical: Develop a customized replanning module for MATSim and an events handler to analyze its effects. 【Schedule】 Aug. 19-23 from 8:45-14:20
教科書・参考書等	Ball, Philip. Critical mass: How one thing leads to another. Macmillan, 2004. Axhausen, Kay Werner, Andreas Horni, and Kai Nagel. The multi-agent transport simulation MATSim. Ubiquity Press, 2016.
授業で使用するメディア・機器等
【詳細情報】
授業で取り入れる学習手法
予習・復習へのアドバイス	Please bring your own laptop in every lecture
履修上の注意受講条件等
成績評価の基準等	Each topic covered will present a small coding challenge. The aggregate grade will be computed from these challenges, each weighted by the number of lectures consumed by the topic
実務経験
実務経験の概要とそれに基づく授業内容
メッセージ
その他
すべての授業科目において，授業改善アンケートを実施していますので，回答に協力してください。回答に対しては教員からコメントを入力しており，今後の改善につなげていきます。

シラバスTOPへ