US 11,836,657 B2
Resource management planning support device, resource management planning support method, and programs
Yuki Maekawa, Tokyo (JP); Tomoe Tomiyama, Tokyo (JP); and Kenichiro Okada, Tokyo (JP)
Assigned to HITACHI, LTD., Tokyo (JP)
Appl. No. 17/642,151
Filed by HITACHI, LTD., Tokyo (JP)
PCT Filed Sep. 10, 2020, PCT No. PCT/JP2020/034261
§ 371(c)(1), (2) Date Mar. 10, 2022,
PCT Pub. No. WO2021/090572, PCT Pub. Date May 14, 2021.
Claims priority of application No. 2019-203593 (JP), filed on Nov. 8, 2019.
Prior Publication US 2022/0327451 A1, Oct. 13, 2022
Int. Cl. G06Q 10/0631 (2023.01); G06Q 50/30 (2012.01)
CPC G06Q 10/06312 (2013.01) [G06Q 10/06311 (2013.01); G06Q 10/06313 (2013.01); G06Q 50/30 (2013.01)] 13 Claims
OG exemplary drawing
 
1. A resource management planning support device that creates an operation rescheduling plan for a train set, the resource management planning support device comprising:
a memory that stores plan result information regarding an existing operation rescheduling plan;
a display; and
a processor communicatively coupled to the memory and the display, wherein the processor is configured to:
exercise a prior learning function to generate duty recommendation models for recommending a duty,
exercise a planning function to create an operation rescheduling plan by selecting at least one of a plurality of duty templates suitable for each of a plurality of planning target train sets according to the duty recommendation models generated,
output a plan result display screen for displaying, via the display, the operation rescheduling plan created,
generate, for each input format of a plurality of input formats designated by a user, a duty learning dataset including a respective input format and at least one of the plurality of duty templates, based on the plan result information stored in the memory,
generate a duty recommendation model including a classifier configured to learn at least one of the plurality of duty templates based on the respective input format from the duty learning dataset,
generate, for each of the plurality of planning target train sets, a train set input value based on an input format of the duty recommendation model from information regarding a corresponding planning target train set,
input the generated train set input value to the classifier included in the duty recommendation model,
create a template priority list of priority levels of each of the plurality of duty templates, the priority levels being obtained as output from the classifier,
preferentially select, based on the template priority list created, one of the plurality of duty templates having a high priority level indicated in the template priority list,
create a plan for each of the plurality of planning target train sets,
create the operation rescheduling plan for each of the plurality of planning target train sets,
output, via the display, a recommendation model selection screen that allows the user to select one duty recommendation model for use with the planning function, from the duty recommendation models generated with use of the prior learning function, and
generate the template priority list by using the duty recommendation model selected from the recommendation model selection screen.