In December 2018, a team of researchers of DeepMind (owned by Google) published a paper in the journal Science, demonstrating the ability of their newly developed AlphaZero algorithm to beat the best game engines in Chess, Go, and Shogi. What is more, instead of relying on handcrafted evaluation functions of board states, the AlphaZero algorithm contains no expert information on any of the games played: it autonomously learns to play each game only by
playing the game many times against itself.

This AI breakthrough is exciting because Go and Chess are games where it is crucial to anticipate unknown moves of the opponent. When making logistics decisions, it is equally important to anticipate the arrival of new data (e.g., orders, delays, disruptions, etc.). For various dynamic data-driven decision problems, Deep Reinforcement Learning (DRL) algorithms like AlphaZero have been demonstrated to be game-changers. The Dutch logistics sector recognizes
the opportunities and is eager to adopt AI for decision automation. However:
• Companies struggle to translate the abstract possibilities of AI into the tangible project plans that are needed to progress towards actual implementations.
• Employing DRL-based decision making requires expert algorithmic knowledge that is difficult to source. Our project goals are to simultaneously overcome these two challenges:
• We develop proofs-of-concept (PoCs) AI decision automation for our 10 industrial
partners, which serve as concrete examples of the potential of AI in data-driven logistics.
• In a similar fashion as AlphaZero was designed as a generic tool to solve various games, we create the DynaPlex toolbox to support the rapid development of automated decision making based on DRL. DynaPlex focuses on dynamic data-driven logistics challenges, and it is crucial in delivering the PoCs for partner companies, while also supporting decision automation for logistics challenges of companies outside the consortium.

Project work packages aim to maximize the synergies of working towards these two goals:
(1) In close collaboration with the companies supporting this consortium, we formalize a wide range of logistics challenges from our 10 industrial partners, which will be incorporated and solved in our framework;
2) We develop a modelling framework that vastly simplifies the process of incorporating new challenges in the toolbox; (3) We develop an algorithmic framework based on DRL that automates solving these challenges; and (4) We develop PoCs and demonstrators to make the potential of AI more concrete for logistics practitioners. We have extensive safeguards in place to ensure that the toolbox will be generally usable.

Our final aim is to support modeling of data-driven logistics decision problems in uncertain
environments utilizing real-time information, and letting the toolbox optimize these problems, with zero coding. This is highly innovative!

Facts & Figures

translations.project.date_start: 1 March 2021 translations.project.date_end: 1 November 2024

Scientific publications

Wetenschappelijk artikel arXiv
wetenschappelijk artikel arXiv
wetenschappelijk artikel SSRN
Wetenschappelijk artikel in Annals of Operations Research
wetenschappelijk artikel Smart Industry-Better Management
wetenschappelijk artikel EJOR
wetenschappelijk artikel conference learning representations
wetenschappelijk artikel Computers & Industrial Engineering
Wetenschappelijk artikel arXiv
wetenschappelijk artikel Transportation Science
conference book Industrial Engineering & Business Information Systems
Conference report - Computational Logistics

Stay up to date with the latest news