1 of 31

Classifications and Generalizations of �Traffic Flow Models in Statistical Mechanics

Department of Physics, Chuo University, Tokyo

1

Mini-Symposium with Prof. Gregory Schehr at Chuo University

Aug 18th, 2023

Saori MORIMOTO

2 of 31

Outline

  • Introduction
  • Modelling by Stochastic Cellular Automata
  • Dependence of Fundamental Diagrams on Basic Models
  • Two Types of New Stochastic Models Interpolating the Basic Models
  • Future Problems

2

3 of 31

Outline

  • Introduction
  • Modelling by Stochastic Cellular Automata
  • Dependence of Fundamental Diagrams on Basic Models
  • Two Types of New Stochastic Models Interpolating the Basic Models
  • Future Problems

3

4 of 31

Introduction

4

▶︎ Problem

▶︎ Purpose

: Physicists have suggested too many distinct models.

: Interpolate the models and discuss how to reduce traffic jam.

5 of 31

Outline

  • Introduction
  • Modelling by Stochastic Cellular Automata
  • Dependence of Fundamental Diagrams on Basic Models
  • Two Types of New Stochastic Models Interpolating the Basic Models
  • Future Problems

5

6 of 31

Modeling Method

6

▶︎ Using Models

▶︎ Boundary Condition

▶︎ Update Rule

: Cellular Automaton Models

: Periodic Boundary Condition

: Parallel Update

7 of 31

Rule-184 CA model

7

 

 

 

 

 

 

 

 

 

 

t steps later

 

 

 

 

 

 

 

 

 

 

t+1 steps later

If the front cell is empty, the car moves forward.

But if the front cell is occupied by another car, the car stops.

This is the simplest model and almost all models are its extensions.

✔︎

✔︎

8 of 31

Quick-Start model

8

The drivers can anticipate the next time step.

Even though the front cell is occupied by another car, if that preceding car is able to move at the next step, then the car moves.

✔︎

✔︎

 

 

 

 

 

 

 

 

 

 

t+1 steps later

expected

t+1 steps later

 

 

 

 

 

 

 

 

 

 

t steps later

9 of 31

Slow-to-Start model

9

When the front cell is empty, the car postpones to move for one time-step.

✔︎

 

 

 

 

 

 

 

 

 

 

t steps later

 

 

 

 

 

 

 

 

 

 

t+1 steps later

 

 

 

 

 

 

 

 

 

 

t+2 steps later

 

 

 

 

 

 

 

 

 

 

t+3 steps later

10 of 31

Outline

  • Introduction
  • Modelling by Stochastic Cellular Automata
  • Dependence of Fundamental Diagrams on Basic Models
  • Two Types of New Stochastic Models Interpolating the Basic Models
  • Future Problems

10

11 of 31

Fundamental diagram

11

density

Flow

Fundamental diagram of empirical data for the CHUO express way in Japan, inbound, Near Sagami Lake

https://www.jstage.jst.go.jp/article/isciesci/63/7/63_271/_pdf/-char/ja

The fundamental diagram is the graph showing the flow of cars versus the density of cars.

✔︎

12 of 31

Fundamental diagram

12

density

Flow

 

 

Fundamental diagram of empirical data for the CHUO express way in Japan, inbound, Near Sagami Lake

https://www.jstage.jst.go.jp/article/isciesci/63/7/63_271/_pdf/-char/ja

13 of 31

Fundamental diagram

13

density

Flow

 

 

Free-flow phase

Fundamental diagram of empirical data for the CHUO express way in Japan, inbound, Near Sagami Lake

https://www.jstage.jst.go.jp/article/isciesci/63/7/63_271/_pdf/-char/ja

✔︎

✔︎

14 of 31

Fundamental diagram

14

density

Flow

 

Jamming phase

Fundamental diagram of empirical data for the CHUO express way in Japan, inbound, Near Sagami Lake

https://www.jstage.jst.go.jp/article/isciesci/63/7/63_271/_pdf/-char/ja

 

✔︎

✔︎

15 of 31

Fundamental diagram

15

density

Flow

 

Metastable states

Fundamental diagram of empirical data for the CHUO express way in Japan, inbound, Near Sagami Lake

https://www.jstage.jst.go.jp/article/isciesci/63/7/63_271/_pdf/-char/ja

 

✔︎

✔︎

16 of 31

Three parts of the fundamental diagram

16

density

Flow

▶︎ The free-flow phase

 

▶︎ The jamming phase

▶︎ The metastable states

The flow in this phase turns to decrease as the density increases.

The flow in this phase is proportional to the density.

 

Free-flow phase

Jamming phase

Metastable states

Fundamental diagram of empirical data for the CHUO EXPWY in Japan, inbound, Near Sagami Lake

https://www.jstage.jst.go.jp/article/isciesci/63/7/63_271/_pdf/-char/ja

17 of 31

Fundamental diagram for Rule-184

17

Number of Cells is L = 100, Initial states is 10 pattern, Time steps is t = 250 ~ 300

 

✔︎

✔︎

 

18 of 31

Fundamental diagram for Quick-Start model

18

✔︎

✔︎

Number of Cells is L = 100, Initial states is 10 pattern, Time steps is t = 250 ~ 300

 

 

19 of 31

Fundamental diagram for Slow-to-Start model

19

✔︎

✔︎

Number of Cells is L = 100, Initial states is 10 pattern, Time steps is t = 250 ~ 300

 

 

20 of 31

Outline

  • Introduction
  • Modelling by Stochastic Cellular Automata
  • Dependence of Fundamental Diagrams on Basic Models
  • Two Types of New Stochastic Models Interpolating the Basic Models
  • Future Problems

20

21 of 31

Two types of new stochastic models

21

Rule-184

SlS

QS

 

I introduce a stochastic rule interpolating the Rule-184 and the QS model.

22 of 31

Two types of new stochastic models

22

Rule-184

SlS

QS

 

 

I introduce a stochastic rule interpolating the Rule-184 and the SlS model.

23 of 31

 

I introduce a stochastic rule :

23

Number of Cells is L = 100, Initial states is 10 pattern, Time steps is t = 250 ~ 300

 

 

 

✔︎

✔︎

24 of 31

 

24

 

 

 

 

25 of 31

 

I introduce a stochastic rule :

25

Number of Cells is L = 100, Initial states is 10 pattern, Time steps is t = 250 ~ 300

 

 

 

✔︎

✔︎

26 of 31

 

26

 

 

 

 

 

27 of 31

Sizes of metastable states

27

Number of Cells is L = 100, Initial states taken average is 1000 pattern, Time steps is t = 250 ~ 300

✔︎

✔︎

I calculated the excess of the maximum flow over the flow at the critical density.

 

maximum flow

flow at the critical density

28 of 31

Summary

28

 

 

 

 

 

29 of 31

Outline

  • Introduction
  • Modelling by Stochastic Cellular Automata
  • Dependence of Fundamental Diagrams on Basic Models
  • Two Types of New Stochastic Models Interpolating the Basic Models
  • Future Problems

29

30 of 31

Future Problems

30

Drivers’ Individual effects

Environmental effects

They include car performances, how to drive, etc.

They include road structures, surrounding facilities, etc.

▶︎ QS model

  ( It deals with the driver’s anticipation effect. )

▶︎ SlS model

  ( It deals with the inertia effect of cars. )

▶︎ TASEP [ Totally Asymmetric Simple Exclusion Process ]

  ( It deals with the random braking effect. )

▶︎ two-lane model

  ( It deals with the lane change effect. )

I want to propose two categories of effects acting on cars

I wish to understand essential mechanism of traffic jam and obtain mathematical expressions with focusing on this classification.

31 of 31

Thank you very much !

31