Analytical method for time allocation of truck freight through machine learning
| Reference number | |
| Coordinator | LUP TECHNOLOGIES AB |
| Funding from Vinnova | SEK 500 000 |
| Project duration | May 2016 - October 2016 |
| Status | Completed |
Important results from the project
This overall purpose of the project is to facilitate the freight buying company´s work processes while to benefit the entire transport chain´s productivity through improved information sharing and better utilization of the logistics chain´s resources. It would be fulfilled with the goal of developing a tool for truck freight buyers, in the form of a machine learning time allocation for inbound- and outbound deliveries by truck. This has led to a working prototype which uses machine learning time allocation and been customer tested.
Expected long term effects
The prototype predicts trucks arrivals given previous arrivals with the help of machine learning and would help companies to better resource planning their operations. It gives better results than statistical methods to predict the arrival times of trucks. The tests show that the learning curve for the prototype guesses accuracy increases with the amount of data. However, the prototype has only been tested on one data set and one cannot make too precise conclusions without verifying the results against other data sets.
Approach and implementation
The work was divided into three steps; needs / requirements specification, development and testing, which was defined as follows: 1. Establish contact with customers and project partners to gather all the necessary information and define the project scope. 2. Develop and finalize the machine learning time allocation and attract customer interest for the implementation of further tests. 3. Test the prototype of the machine learning time allocation at a customer and analyze the results