HYBRID SYSTEM AND METHOD FOR PRODUCING OPTIMIZED TRAINING PROGRAM SCHEDULES

Patent №

US 11,295,403

Granted

2022-04-05

Filed 2018

Owner

BRITANNICA KNOWLEDGE SYSTEMS LTD.

+1 more

Lab

AI components

4

ml · planning · evo · hardware

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

16170669

A hybrid computer-based system and method for scheduling a plan of training programs, such as flight training programs, is disclosed. A heuristic scheduler receives a plan request comprising training programs to schedule, along with a plan vector, scheduling rules, and available resources. The scheduler computes a locally optimized scheduling plan of the training schedules in the request. An evaluator evaluates one or more costs of the scheduling plan. A solver employs stochastic methodology of reordering the plan vector, in order to improve the costs. The scheduler, evaluator, and solver are employed iteratively until a globally optimized scheduling plan is determined. In some embodiments, scheduling rules are arranged in a sequence of constraint-relaxed tiers, allowing for flexibility in heuristic scheduling in light of multiple layers of constraints: validity, i.e., inviolable (hard) constraints that training schedules must meet.

AI classification

Machine learning1.00
Planning1.00
Evolutionary computation0.99
AI hardware0.76
Knowledge representation0.12
Natural language0.00
Vision0.00
Speech0.00

Ownership

BRITANNICA KNOWLEDGE SYSTEMS LTD.

assignment · 499210697

QUALTERO LTD

namechg · 639820900

Assignors

RISSIN, YOSSEF, DAGAN, ASHER, BEN-ISHAY, YOSEF

On an employer assignment, the assignors are typically the inventors.

© 2026 NYSGPT2525 LLC