AIST and JR-East

1. Introduction

RW-Solver is a formal tool for generating train timetables with the help of SMT-Solver, in a mesoscopic level that is more concrete than the macroscopic level, where each station is simply expressed in a black-box, and more abstract than the microscopic level, where the infrastructure in each station-area is expressed in detail. The accuracy of generated timetable and the computational effort for the generation is a trade-off.

About the accuracy of railway models, the following three levels are often discussed:

- In the microscopic level, models can express the details of the railway networks and the platforms in each station area as shown in Fig.1(a). In this level, train timetables are often improved by simulations.
- In the macroscopic level, models usually represent each station as a black-box as shown in Fig.1(c), and schedules in the models consist of departure and arrival times at the black-box.
- Mesoscopic levels are more concrete than the macroscopic level and more abstract than the microscopic level. For example, in a mesoscopic level models can express available platforms for each direction in each station area as shown in Fig.1(b) and it is still possible to automatically generate timetable even if it is difficult in the microscopic level.

Fig.1 Micro-, meso-, and macroscopic models of a station

About the strategy for solving constraints, there are the following two classes of methods:

- Exact solution method: Constraint solvers such as SAT-Solver are used for checking whether a timetable to satisfy the given set of constraints exists or not. Such a problem is known to be NP-complete.
- Heuristic solution method: Some heuristic approaches such as genetic algorithm and greedy procedure have been applied for heuristically generating timetables to satisfy the given set of constraints in reasonable computational time.

Fig.2 The overview of the generation of timetables by RW-Solver

2. License

- Following the ideas of open source software, we allow anyone to use RW-Solver without fee, under the GNU General Public License version 3 (GPL v3). You have to agree with the license before using RW-Solver.

- The OCaml source files of RW-Solver can be download from this GitHub
site.

- AIST Program-IP No. H30PRO-2219
- National Institute of Advanced Industrial Science and Technology (AIST) and

East Japan Railway Company (JR-East)

- The manual for RW-Solver [English: PDF, Japanese: PDF]

5. Publication

- Yoshinao Isobe, Hisabumi Hatsugai, Akira Tanaka, Yutaka Oiwa,
Takanori Ambe, Akimasa Okada, Satoru Kitamura, Yamato Fukuta, and
Takashi Kunifuji, Automatic Generation of Train Timetables from
Mesoscopic Railway Models by SMT-Solver, IEICE
Transactions on Fundamentals of Electronics, Communications and
Computer Sciences, Vol.E102-A, No.2, pp.325-335, 2019. [PDF
(5.5MB)]

Yoshinao Isobe

Cyber Physical Security Research Center

National Institute of Advanced Industrial Science and Technology (AIST), Japan

Cyber Physical Security Research Center

National Institute of Advanced Industrial Science and Technology (AIST), Japan

Last modified: 7 February 2019

(since 07/02/2019)