TY - JOUR
T1 - Development of reactive scheduling for rolling stock operation using a constraint model
AU - Tomiyama, Tomoe
AU - Sato, Tatsuhiro
AU - Murata, Tomohiro
AU - Iwamura, Shigeki
AU - Sakamoto, Osamu
PY - 2017
Y1 - 2017
N2 - In almost all Japanese railway companies, rolling stock management still relies on manual procedure. As density of trains and interconnectivity between lines are increased, a support system for the management gets to be important. This paper focuses on a scheduling task for vehicle units' operation and a support system for rescheduling vehicle units' operation in train service disruption. A schedule of vehicle units' operation is modified by a dispatcher to keep consistency between train timetable and vehicle unit assignment. When there is no solution, the dispatcher has to ask a train service manager to change train timetable. To support this task, we developed an algorithm of rescheduling vehicle unit operation, which included relaxation of constraints. Our algorithm represents constraints as a network model, and, if necessary, relaxes constraints by modifying the network model during search of a solution. This paper indicates the network model, the algorithm of searching for a solution as relaxing constraints, and numerical experiments using a real middle size data. The results of the experiments showed that our algorithm could calculate various solutions with relaxed constraints. Analyzing the solutions at the view points of train service and rolling stock operation, we showed the effectiveness of a system that proposes various solutions using our algorithm.
AB - In almost all Japanese railway companies, rolling stock management still relies on manual procedure. As density of trains and interconnectivity between lines are increased, a support system for the management gets to be important. This paper focuses on a scheduling task for vehicle units' operation and a support system for rescheduling vehicle units' operation in train service disruption. A schedule of vehicle units' operation is modified by a dispatcher to keep consistency between train timetable and vehicle unit assignment. When there is no solution, the dispatcher has to ask a train service manager to change train timetable. To support this task, we developed an algorithm of rescheduling vehicle unit operation, which included relaxation of constraints. Our algorithm represents constraints as a network model, and, if necessary, relaxes constraints by modifying the network model during search of a solution. This paper indicates the network model, the algorithm of searching for a solution as relaxing constraints, and numerical experiments using a real middle size data. The results of the experiments showed that our algorithm could calculate various solutions with relaxed constraints. Analyzing the solutions at the view points of train service and rolling stock operation, we showed the effectiveness of a system that proposes various solutions using our algorithm.
KW - Constraint programming
KW - Rescheduling vehicle unit operation
KW - Rolling stock operation
UR - http://www.scopus.com/inward/record.url?scp=85026526997&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85026526997&partnerID=8YFLogxK
U2 - 10.1541/ieejeiss.137.1121
DO - 10.1541/ieejeiss.137.1121
M3 - Article
AN - SCOPUS:85026526997
VL - 137
SP - 1121
EP - 1131
JO - IEEJ Transactions on Electronics, Information and Systems
JF - IEEJ Transactions on Electronics, Information and Systems
SN - 0385-4221
IS - 8
ER -