2004
Initial Idea for Swarm Logistics
- During his studies at the University of Tübingen, focusing on Quantitative Finance and Game Theory, founder Damir Dulovic conceived the first idea for Swarm Logistics.
- Growing up in the logistics sector and running his own transport company to finance his studies provided practical insights into the logistics industry.
- A lecture on Algorithmic Trading inspired the idea to use autonomous agents in logistics, similar to the financial sector using trading bots.
- These agents were envisioned to evaluate and plan, automatically accept orders, negotiate fleets among themselves, and subsequently execute tasks.
- However, technology at the time was not sufficiently advanced; communication was mainly via SMS, often without mobile internet, and there were no workstations, SaaS, or cloud solutions.
- Initial considerations of Reverse Game Theory—optimization through negotiation.
2007
Recognition of Mechanism Design
- The Nobel Prize in Economic Sciences was awarded in 2007 to Leonid Hurwicz, Eric Maskin, and Roger Myerson “for having laid the foundations of mechanism design theory.”
- This recognition provided the theoretical basis for applying Mechanism Design in logistics.
2015
Technology Enables the Vision
- Technology had advanced enough to implement the vision.
- Daimler tested the first self-driving truck on the road.
- It became clear that someone needed to coordinate these vehicles.
- At that time, dispatching was almost entirely manual.
- Delivery notes were printed and used on a table for dispatching, often with fixed delivery zones.
- Cross-fleet collaboration was conducted via clipboard, as were freight exchanges.
Concept Development
- Trucks coordinate like swarms with flexible times and locations for transshipment, form into platoons, and continuously reorganize.
- Economic agents evaluate routes and freights, calculate freight rates, plan, and dynamically negotiate these with other agents based on an adapted cost accounting system.
- Coordination occurs both within and across fleets through negotiations based on Mechanism Design using a mediator; local problem-solving, global optimization.
- Operations Research-based optimizations require complete information sharing among all participants and are not dynamically solvable for very large fleets, even with heuristics; multi-agent systems are.
2016
February
Co-Founder Harry Trautmann
- Met at a conference at CyberForum.
- Harry studied under Prof. Homberger, one of the world’s most respected researchers in multi-agent systems, and his thesis focused on multi-agent systems, with his algorithm ranking top worldwide for years.
Expansion of the Vision
- Aimed to achieve a machine economy in logistics and become the first company to create vertical integration.
- Develop the first self-organized company with an agent-based robo-manager, then referred to as a Business Operating System, which makes decisions for a profit center.
April
Ethereum Presents Smart Contract-Based Decentralized Autonomous Organization (DAO)
- Until then, considerations were only on the logical level; with smart contracts, there was now the possibility of a physically decentralized version.
- Perfect match between decentralized multi-agent systems and blockchain.
- Blockchain provides a physical layer allowing participants who don’t know each other to collaborate without intermediaries, generating trust.
- Distinction between DAO and DAC based on the type of smart contract, i.e., cross-company or intra-company with different mediators and objective functions.
- The mediator doesn’t have to be a central intermediary but can also run as a smart contract on a blockchain platform.
2017
First Version of an Auto-Dispatcher
- Created a desktop version of the auto-dispatcher based on multi-agent systems.
- Developed the first version of autonomous agents that evaluate freights, plan, and negotiate with other agents.
- First pilot project with flexible buses for patient transport and disabled children.
Technical Evaluation
- Initial technical evaluation of Blockchain and Distributed Ledger Technology (DLT).
2018
Founding of Swarm Logistics GmbH
- Official company founding to implement the developed concepts.
- Extensive applied research and development.
2019
Admission to the ESA BIC Incubation Program
- Joined the European Space Agency’s (ESA BIC) incubation program with a grant.
2020
Laboratory Prototype of a Fully Decentralized Network
- Developed a prototype with mini-PC-based simple agents and IOTA nodes.
- Fully serverless infrastructure.
World’s First Vertical Integration of a Machine Economy
- World’s first implementation of a machine economy in an industrial sector with autonomous agents collaborating over a blockchain.
- Finalist at the Cyber One Hightech Award.
2021
Seed Investment from ETO Group
- Expanded the team.
Patent Applications
- Filed patents in Germany and Europe.
Development of the Basis for Agents
- Adapted a cost accounting system based on relative direct costs according to Prof. Riebel for freight rate calculation.
- Extended with multi-cost factors and adjustments for offers.
- Combined and extended with route planning logics to minimize costs.
- Forked and extensively customized an open-source routing engine for evaluations in the tour.
2022
Field Tests
- Confirmed the use of multi-agent systems even in static route planning environments.
- Necessity for precise route durations and distances for realistic cost calculations.
- Extended the MinCost routing engine with machine-learning functions for extremely precise ETA predictions and cost calculations with outstanding success.
Challenges
- Impact of COVID-19 and reluctance among OEMs with long development cycles for onboard units.
Software-Defined Vehicles (SDV)
- SDVs became mainstream; the agent could now operate as a full software agent in the application layer of an SDV.
Strategic Decision
- Developed a Carrier & Navigation App equipped with the agent to enable pilot projects in the field and serve as a navigation replacement in vehicles.
Admission to Cyber Valley
- Joined the startup network of Cyber Valley.
2023
Development of the MyT.One Carrier App – Mighty App
- Finalized the economic agent code.
- Functions: Evaluate, Plan, Execute.
Beta Version of the Auto-Dispatcher
- Provided a cloud infrastructure and API for TMS providers with complete peripheral systems.
- Shippable intermediary products with underlying visionary technology.
2024
Market-Ready Rollout of an Automatic Dispatcher
- Via REST API based on multi-agent systems as an intermediate product, logically decentralized but still physically centralized.
- Logical level is in use and functional.
Patent Applications in the USA, China, and India
Product Development
- Rolled out standalone products that are used internally by the auto-dispatcher.
Successful Pilot Project with Telematics Provider Arealcontrol GmbH, Stuttgart
- ETA prediction with over 6,500 routes.
- Validation of results: prediction precision about 3%–10%, in the minute range over very long distances—world-class results and best in class for trucks in Europe.
Introduction of Additional Agents
- Integrated open-source LLMs in matrix multi-agent structures with heterogeneous agents.
- Supported by expert systems like routing and ML-based ETAs for evaluations.
- Deployed additional agents for interpreting documents, evaluating cargo space and loads, and assessing available capacities.
Achievements
- Reduced dispatching costs by 15–35%.
- Shortened planners’ planning time by up to 95%—from hours to minutes.
- Over 100 installations.
- To achieve vertical integration for an industry sector, dozens of technologies had to be addressed and extremely difficult individual problems had to be solved.
2025 (Planned)
Planned Field Tests and Migration
- Migration of agents and mediators to a DLT for the world’s first machine economy implementation on DLT/Blockchain.
- End-to-end vertical integration of a Decentralized Autonomous Logistics Organization (DALO) in productive use.
Extension of Smart Contracts
- Support for competing fleets and smart contracts for Decentralized Autonomous Corporations (DAC).
- Extension for revolving planning around the clock without human intervention.
Integration into SDVs
- Economic agents available in the application layer at OEMs.
Networking with Other Agents
- Connection with agents of forklifts, intelligent ramps, LLMs, and agents that control other corporate areas.
Machine Economy Achieved!
Vision
A self-organizing network of vehicles without central authority and control. Collaboration of competing vehicles and fleets over a blockchain and serverless infrastructure without intermediaries. The economic AI agent as a robo-manager is installed in the vehicle, makes autonomous and private decisions, and negotiates with other economic agents of any cyber-physical systems (e.g., forklifts, production machines). Cross-company collaboration while maintaining data and decision sovereignty.
This is part of the vision of a self-organizing economy with collaboration and utilization of additional agents as part of the Machine Economy/Metaverse Economy/Metamobility. A technological upgrade for existing companies that collaborate on equal footing, based on a resilient network protected against cyber-attacks and maintaining sovereignty over central or monopolistic platforms, where the network belongs to the participants. A European way.
Vehicles coordinate like swarms.