Machine Learning Model for Vehicle Task Assignment

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for assigning vehicles to tasks in ride hailing and ride sharing services face computational challenges, making it difficult to find optimal assignments efficiently, especially when considering factors like vehicle availability, maintenance, and charging status.

Innovation Solution

A machine-learning model is used to evaluate and assign vehicles to tasks by providing information on vehicles and tasks as input, allowing for quick evaluation of assignments and selecting the most favorable options based on an objective function, which can include factors like waiting time, vehicle wear, and user ratings.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional computational methods are used to optimize vehicle assignment, then assignment optimality can be achieved, but computational effort and time consumption increase significantly

Engineering Contradiction:
Improveassignment optimalityVSAvoidcomputational time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system pre-trains machine learning models using historical assignment data and simulation environments before actual vehicle assignment occurs. This preliminary training enables the model to quickly evaluate assignments during operation without requiring intensive real-time computation, thus achieving optimal assignments with reduced computational time.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

A machine learning model serves as an intermediary between the complex optimization problem and the final assignment decision. The model learns to approximate the optimal assignment function from training data, acting as a mediator that translates vehicle and task features into favorable assignments without requiring exhaustive computational search during deployment.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If comprehensive factors (waiting time, vehicle wear, user ratings) are considered in assignment optimization, then assignment quality improves, but computational complexity increases

Engineering Contradiction:
Improveassignment qualityVSAvoidcomputational complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system transforms the complex multi-factor optimization problem into a machine learning prediction task by changing the parameters from direct optimization variables to input features. The model takes vehicle state, task characteristics, and historical data as inputs and directly outputs favorable assignments, avoiding the need to explicitly optimize multiple conflicting parameters.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

Traditional mechanical optimization algorithms (linear programming, integer programming) are replaced with a data-driven machine learning approach. Instead of iteratively solving complex mathematical programs, the pre-trained model directly predicts optimal assignments by learning patterns from historical data, substituting computational mechanics with statistical inference.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP3806007B1Methods, computer programs and systems for assigning vehicles to vehicular tasks and for providing a machine-learning model
Publication Date: 2024.08.28 BAYERISCHE MOTOREN WERKE AG
  • EP3806007B1 patent drawingFigure 1a~1c
  • EP3806007B1 patent drawingFigure 2a~2b
  • EP3806007B1 patent drawingFigure 2c~4

AI summary

Examples relate to a method, computer program and system for assigning a plurality of vehicles to a plurality of vehicular tasks and to a method, computer program and system for providing a machine-learning model. The method for assigning a plurality of vehicles to a plurality of vehicular tasks comprises obtaining information on the plurality of vehicular tasks. The method comprises obtaining information on the plurality of vehicles. The method comprises providing the information on the plurality of vehicular tasks and the information on the plurality of vehicles as input to a machine-learning model. The machine-learning model is suitable for evaluating a plurality of assignments between the plurality of vehicles and the plurality of vehicular tasks. The method comprises assigning the plurality of vehicles to the plurality of vehicular tasks based on an output of the machine-learning model.