# US 11,077,873 B2 · System and method for train route optimization including machine learning system > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,077,873 B2 - Document ID: US11077873B2 - Status: Granted - Kind: B2 - Application number: 16/309,152 - Filed: 2017-06-13 - Published: 2019-08-15 - Granted: 2021-08-03 - Earliest claimed: 2016-06-13 - Assignees: Siemens Mobility GmbH - Applicants: Siemens Industry, Inc. - Inventors: Badar K. Khan; Michael Golm; Michael Swieczkowski - Web page: https://patentgrep.com/patent/US11077873B2 ## CPC - B61L 27/16: Trackside optimisation of vehicle or train operation - B61L 15/0058: {On-board optimisation of vehicle or vehicle train operation} - B61L 25/021: {Measuring and recording of train speed} - B61L 25/025: {Absolute localisation, e.g. providing geodetic coordinates} - B61L 25/04: Indicating or recording train identities - B61L 27/12: Preparing schedules - G01C 21/3446: {Details of route searching algorithms, e.g. Dijkstra, A*, arc-flags or using precalculated routes} - G06F 7/00: Methods or arrangements for processing data by operating upon the order or content of the data handled (logic circuits H03K19/00) - B61L: GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC (brakes or auxiliary equipment B61H, B61K; point or crossing construction E01B) - G01C: MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY (measuring liquid level G01F; radio navigation, determining distance or velocity by use of propagation effects, e.g. Doppler effects, propagation time, of radio waves, analogous arrangements using other waves G01S) - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract A machine learning system (100) for train route optimization includes a machine learning module (120) in communication with an optimization module (110) and including instructions stored in a memory that when executed by a processor (82) cause the machine learning module (120) to receive a plurality of schedules for railroad vehicles travelling through a track network transmitted by the optimization module (110), prioritize the plurality of schedules by applying at least one business rule (140) to the plurality of schedules based on dynamically learned behavior, and provide a prioritized list of schedules. ## Claims ### Claim 1 A machine learning system for train route optimization comprising: a machine learning module in communication with an optimization module and including instructions stored in a memory that when executed by a processor cause the machine learning module to: receive a plurality of schedules for railroad vehicles travelling through a track network transmitted by the optimization module, prioritize the plurality of schedules by applying at least one business rule to the plurality of schedules based on dynamically learned behavior, and provide a prioritized list of schedules, wherein the optimization module is configured to generate the plurality of schedules by optimizing initial schedules utilizing a genetic algorithm, wherein the initial schedules comprise at least one schedule with delays of one or more railroad vehicles travelling in the railroad network. ### Claim 2 (depends on 1) The machine learning system of claim 1, wherein the dynamically learned behavior is based on instructions and decisions provided by a human operator. ### Claim 3 (depends on 1) The machine learning system of claim 1, further comprising: a human-machine-interface (HMI) configured to display the prioritized list of schedules provided by the machine learning module, and to permit a user to modify the prioritized list of schedules in view of the at least one business rule. ### Claim 4 (depends on 3) The machine learning system of claim 3, wherein the instructions, when executed by the processor, further cause the machine learning module to: modify the prioritized list of schedules based on instructions received from the user via the HMI. ### Claim 5 (depends on 1) The machine learning system of claim 1, wherein the at least one business rule includes mitigation of track conflicts of one or more railroad vehicles travelling on a same railroad track. ### Claim 6 (depends on 1) The machine learning system of claim 1, wherein the instructions, when executed by the processor, further cause the machine learning module to: analyze the plurality of schedules by evaluating a number of conflicts removed from each schedule. ### Claim 7 (depends on 1) The machine learning system of claim 1, further comprising: a user mode input comprising an expert user mode and a novice user mode, wherein the dynamically learned behavior of the machine learning module is based on input and instructions provided during expert user mode. ### Claim 8 (depends on 7) The machine learning system of claim 7, wherein, during the novice user mode, the machine learning module is configured to prioritize the plurality of schedules based on a training set of expert data stored within the machine learning module. ### Claim 9 A method for optimizing a schedule for railroad vehicles travelling within a railroad network performed by a machine learning system, through operation of at least one processor, comprising: receiving, by a machine learning module in communication with an optimization module, a plurality of schedules for railroad vehicles travelling through a track network transmitted by the optimization module, prioritizing, by the machine learning module, the plurality of schedules by applying at least one business rule to the plurality of schedules based on dynamically learned behavior, and providing, by the machine learning module, a prioritized list of schedules, and further comprising generating, by the optimization module, the plurality of schedules using a genetic algorithm, comprising: providing an initial population of initial schedules, each schedule comprising first information of a railroad track network and second information of at least one railroad vehicle travelling in the railroad track network for a time instance, selecting at least two schedules of the initial schedules, and generating a final population of final schedules utilizing crossover operation between selected initial schedules, wherein the second information of the at least one railroad vehicle travelling in the railroad track network are exchanged between the selected initial schedules. ### Claim 10 (depends on 9) The method of claim 9, wherein the first information of the railroad track network includes one or more parameters comprising identification of track component, track name, track length, track occupied/unoccupied, or cost of track usage and wherein the second information of the railroad vehicles includes one or more parameters comprising identification of the at least one railroad vehicle, train name, train speed, train length, travelling time, delay time, or train priority. ### Claim 11 (depends on 9) The method of claim 9, wherein the initial population comprises at least one schedule with delays of one or more railroad vehicles travelling within the railroad track network. ### Claim 12 (depends on 9) The method of claim 9, further comprising: providing a human-machine-interface configured to display the prioritized list of schedules provided by the machine learning module, and to permit a user to modify the prioritized list of schedules in view of the at least one business rule. ### Claim 13 (depends on 9) A non-transitory computer readable medium comprising computer instructions capable of being executed in at least one processor and performing the method as claimed in claim 9.