What we can do and how? Lingyun Meng, Ph.D., Associate Professor - - PowerPoint PPT Presentation

▶

Mar 15, 2023 123 likes •391 views

Integrated railway operations planning: What we can do and how? Lingyun Meng, Ph.D., Associate Professor State Key Laboratory of Rail Traffic Control and Safety Beijing Jiaotong University, Beijing, China Email: lymeng@bjtu.edu.cn April, 2016

SLIDE 1

Lingyun Meng, Ph.D., Associate Professor State Key Laboratory of Rail Traffic Control and Safety Beijing Jiaotong University, Beijing, China Email: lymeng@bjtu.edu.cn April, 2016

Integrated railway operations planning: What we can do and how?

SLIDE 2

SLIDE 3

April 2016, Dagstuhl Seminar, Germany Many kinds of plans- Line plan in the Netherlands

SLIDE 4

April 2016, Dagstuhl Seminar, Germany

Many kinds of plans-Train timetable for Beijing-Shanghai HSR

SLIDE 5

April 2016, Dagstuhl Seminar, Germany Many kinds of plans-Rolling stock circulation plan in the Netherlands Source: Alfieri, A., Groot, R., Kroon, L., Schrijver, A., Efficient Circulation of Railway Rolling Stock, 2006.

SLIDE 6

April 2016, Dagstuhl Seminar, Germany Many kinds of plans-Crew schedule Source: http://www.aviationsystems.co.uk/wp-content/uploads/2014/03 /Roster-Graphics.jpg

SLIDE 7

Line planning Train routing and Scheduling Vehicle scheduling Crew scheduling Traditionally

April 2016, Dagstuhl Seminar, Germany

SLIDE 8

Question #1: Is it necessary to do integration?

April 2016, Dagstuhl Seminar, Germany

1.1 Does practical work need integration?

My own experience in a dispatching center of one Chinese railway bureau tells me YES! A train dispatcher needs to frequently communicate with rolling stock dispatcher to make sure his dispatch plan is FEASIBLE.

SLIDE 9

Question #1: Is it necessary to do integration?

April 2016, Dagstuhl Seminar, Germany Anita Schöbel: An integrated solution is always as least as good as optimizing stage by stage. (Oct, 2014) Existing studies in integrated line planning train timetabling : Such as: Schöbel and Schmidt (2014), Niu and Zhou (2015), Espinosa-Aranda et al. (2015)

Solution space Hierarchical Integrated

In theory: integration seems to be promising for better solutions

SLIDE 10

Question #2: What are possible topics for us?

April 2016, Dagstuhl Seminar, Germany

Line planning Train routing and Scheduling Vehicle scheduling Crew scheduling Possible integration?

SLIDE 11

Question #2: What are possible topics for us?

April 2016, Dagstuhl Seminar, Germany

Line planning Train routing and Scheduling Vehicle scheduling Crew scheduling Possible integration?

SLIDE 12

Question #2: What are possible topics for us?

April 2016, Dagstuhl Seminar, Germany

Line planning Train routing and Scheduling Vehicle scheduling Crew scheduling Possible integration?

SLIDE 13

Question #2: What are possible topics for us?

April 2016, Dagstuhl Seminar, Germany

Line planning Train routing and Scheduling Vehicle scheduling Crew scheduling Possible integration?

SLIDE 14

Question #2: What are possible topics for us?

April 2016, Dagstuhl Seminar, Germany

Line planning Train routing and Scheduling Vehicle scheduling Crew scheduling Possible integration?

SLIDE 15

Question #2: What are possible topics for us?

April 2016, Dagstuhl Seminar, Germany

Line planning Train routing and Scheduling Vehicle scheduling Crew scheduling

Possible integration?

SLIDE 16

Question #2: What are possible topics for us?

April 2016, Dagstuhl Seminar, Germany

2.1 Are there any interesting integration topics? 2.2 Are there any integration-related research topics? Someone doubted already about this and listed similar methods (for finding research topics) in job shop scheduling research area. Thus we need to find integration topics with real demand from practice and research point of view.

SLIDE 17

Question #3: Challenges if we do integration?

April 2016, Dagstuhl Seminar, Germany 3.1 How complex it will be if we optimize a joint problem?

NP1 NP2 +

Still NP? More complex? 3.2 How can we guarantee that we can find better solutions in a larger solution space efficiently?

Hierarchical Integrated

SLIDE 18

3.3 Does “integration” mean “a holistic framework with sub-models which are solved iteratively ”?

Question #3: Challenges if we do integration?

April 2016, Dagstuhl Seminar, Germany 3.4 Integrated models normally yield in decision variables with high dimensions, then how do we deal with curse of dimensionality? 3.5 If we build an integrated model and we want to do decomposition, how? In a different way from iterative approach.

SLIDE 19

Question #4: How do we meet the challenges?

April 2016, Dagstuhl Seminar, Germany 4.1 Can solution algorithms with parallel computing techniques can solve integrated problems to satisfactory solutions efficiently? 4.2 How much time do we need before we see some substantial processes? 2 years, 5 years, or longer? We need domestic and international collaboration to do integration.

SLIDE 20

SLIDE 21

Ongoing research topics

1. Demand-oriented vehicle routing scheduling for a

transport system Xuesong Zhou, Arizona State University

2. Integrated railway traffic control and train control

Francesco Corman, Xiaojie Luan, TUDelft Yihui Wang, Beijing Jiaotong University

April 2016, Dagstuhl Seminar, Germany

SLIDE 22

April 2016, Dagstuhl Seminar, Germany

Ongoing research topics

1. Demand-oriented vehicle routing scheduling for a

transport system

 Based on a state-space-time network  An integer programming model with vehicle and infrastructure capacity constraints  Solved by Lagrangian relaxation and dynamic programming

SLIDE 23

A special issue on Part C journal about integrated approaches

SLIDE 24

Expected topics in the special issue

The topics of interest include, but are not limited to, the following:

Demand-oriented line planning and train timetabling
Integrated optimization of train timetabling, rolling stock and crew

scheduling

Streamlined train operations planning, infrastructure and rolling stock

maintenance scheduling

Traffic control and train control for demand-oriented urban transit
Train, rolling stock and crew real-time rescheduling.

Welcome to submit a paper to the issue if you have one and no destination journal decided Deadline is May 1st

SLIDE 25

INFORMS RAS Problem Sovling competition

https://www.informs.org/Community/RAS/Problem-Solving- Competition Integrating train timetabling and maintenance task scheduling To be announced soon! Work together with Dr. Francesco Corman from TUDelft

Line_1 Line_2 Line_1

1 3 13 23 26 42 55 44 24 25 33 30 32 31 2 12 49 54 43 50 45 52 53 41 56 58 47 57 20 5 9 4 8 21 6 7 18 19 down up up down up down up down down up 10 11 14 15 17 16 37 39 40 48 46 51 22 27 28 29 34 35 38 36

Line_2

SLIDE 26

Lingyun Meng, Ph.D., Associate Professor State Key Laboratory of Rail Traffic Control and Safety Beijing Jiaotong University, Beijing, China Email: lymeng@bjtu.edu.cn April, 2016