Software agents arrange freight more efficiently

Traffic jams regularly bring Dutch highways to a halt. Freight is often a major cause of such bottlenecks. Interestingly enough almost one third of all trucks drive around empty. PhD student Valentin Robu of the Centrum Wiskunde & Informatica (CWI), the national research center for mathematics and computer science in the Netherlands, examined the efficiency in logistics and saw room for improvement through the use of intelligent software agents.

Publication date: 01-07-2009

Traffic jams regularly bring Dutch highways to a halt. Freight is often a major cause of such bottlenecks. Interestingly enough almost one third of all trucks drive around empty. PhD student Valentin Robu of the Centrum Wiskunde & Informatica (CWI), the national research center for mathematics and computer science in the Netherlands, examined the efficiency in logistics and saw room for improvement through the use of intelligent software agents.


Promotion

On Thursday, 2 July Robu defended his PhD thesis 'Modeling Preferences, Strategic Reasoning and Collaboration in Agent-Mediated Electronic Markets' at the Eindhoven University of Technology. The ultimate goal of the research is to achieve reductions in cost, traffic jam and carbon dioxide emissions.

Automation
In the daily practice of freight transportation, human planners strive to make a return schedule to combine as many loads as possible. This approach is time consuming and often not optimal. During the execution of the trip many things may change: new loads, traffic jams, and so on. Valentin Robu looked at how these environments could be automated with intelligent agents. This independent software programmes can bid on various loads on behalf of their owners. The scientist developed techniques for negotiations and auctions. A decentralized system with intelligent agents appeared to be much faster and better able to adapt to unforeseen circumstances than centralized planning methods. This study was part of the project Distributed Engine for Advanced Logistics (DEAL).

Valentin Robu also researched collaborative tagging techniques. These allow users to collectively find and classify information. Keywords (tags) can be annotated to documents, photos or web pages. For example, on a webpage promoting holidays in Milan people can add words like 'beautiful', 'hot', 'culture' or 'fun city'. Robu looked at whether these tags could be classified into a useful system with a lowest common denominator. He did this with the same techniques he developed for negotiations.

The DEAL project was funded by SenterNovem. Participants were: Almende BV, Erasmus University Rotterdam, Vrije Universiteit Amsterdam, CWI, CarrierWeb, Post-Kogeko and Vos Logistics. At CWI logistics is an important issue. The DEAL project is a good example.

More information:
Valentin Robu, e-mail Valentin.Robu @ cwi.nl. Supervisor: Prof. Dr. Han La Poutre (CWI and TU / e), e-mail hlp@cwi.nl, tel 020-592 4082. Picture: Vos Logistics.