Join our team! Read here what we offer to our members and in which fields of activity you can participate.
Interested? Write an e-mail or visit us at Berliner Tor 7 in room 4.80D.

Requirements

  • Presence: The group has scheduled a monthly meeting whose regular attendance is mandatory. Besides, a frequent presence in our room is desired.
  • Spare time: MARS isn’t complicated – but it is quite complex! You’ll have to take some time to get familiar with the architecture and the tools we use.
  • Commitment: Though everyone has his subject, there is plenty of work to do – from integration and quality assurance to workflow streamlining and testing of new tools. Do your voluntary part!
  • Professionalism: You won’t develop a self-standing application but a part of our framework. Most likely others will depend on your work, thus a reliable and firm way of working is imperative.
  • Active communication: With MARS being a project with over a dozen participants, a lively information exchange is vital for all members.

 

Incentives

  • Workplaces: We offer desks and MacBook Pros with dual-monitor setup for active members.
  • Business-like workflow: A software development, integration and deployment process like in professional companies.
  • State-Of-The-Art tools: MARS uses a multitude of modern and interleaved tools to ease the development. Gitlab, Confluence and various other services are hosted in our infrastructure.
  • Collaboration & research: The MARS group participates in international research projects and is actively publishing papers.
  • A helpful team: Whether you have study or technology-related questions: The MARS team will provide you with help and advice.

 

Themes: Modeling

Core aspects:

  • Trans-Disciplinary Multi-Agent Based Modeling
  • Urban Modeling
  • Urban Logistics
  • Bring Your Own
    • Besides the themes listed hereafter, we also encourage you to join us with your own modeling interests and ideas!
  • Urban Simulation
    • The MARS Urban simulation project aims to model and simulate the traffic flow of bigger cities. This is done with an multimodal approach, meaning different types of traffic participants which interact and influence each other.
      The model strives to demonstrate the large scale capabilities of the MARS framework and relies on Open Data sources like OpenStreetMap. The Project is in an early stage and the first goal is to model the traffic system of Hamburg.
    • Through our international partners, a special focus resides on Smart Urban Logistics
  • ARS AfricaE Model Development

    • Savannah ecology and ecosystem resilience in the Kruger National Park
  • Pedestrian Dynamics & Evacuation

    • Aircraft evacuation
    • Evacuation route planning for huge events (LoveParade etc.)

 

Themes: Infrastructure & Technology

  • Development of a MARS Archive Service:

    Simulation results output needs to be super-fast to not hinder the execution of the Simulation itself. While we do have fast storage solutions, these are not large enough to store all simulation results for a longer period. Precious results from previous simulations therefore need to be archived on larger, but slower storage. In order to ensure integrity, consistency and availability of that data, we need an Archive Service which allows the users of MARS to selectively archive simulation results to a well-defined format and to restore these results for further analysis in the future.
    Technologies: MongoDB, Cassandra, Blob Storage, Microservices, Kubernetes + Docker (only as user)
    Locality: Due to the tight integration requirements with our services, this thesis requires you to work in the MARS lab for at least 2-3 days a week. This is necessary to properly sync with your fellow students and get support from them 😉
    Kind: Bachelor Thesis, if strong sense for quality and good documentation are applied.

  • Big Data Management in Distributed Simulations

    Context: When running simulations, a lot of data needs to be recorded as result set. The various goals such as visualization, analysis and contextualization require different output formats, but should rely on a uniform data stock base for executed simulations, which then are accessed by the adapters tailored for the effort. This thesis is about creating an efficient storage concept for simulation results and implementing that solution together with the MARS team. Among the issues are fast data logging during simulation execution, setting an optimal data format for above goals, provide efficient data access to the results (caching, indexing, etc.) and to provide interfaces and specifications for integration with the MARS framework.
    Technologies: Specialized DBMS (e.g.: VoltDB), Caching Services (memcached etc.), Microservices, Docker
    Locality: Due to the tight integration requirements with our services, this thesis requires you to work in the MARS lab for at least 2-3 days a week. This is necessary to properly sync with your fellow students and get support from them 😉
    Kind: Master Thesis; also Bachelor Thesis, if strong sense for quality and good documentation are applied or it is planned to continue with master thesis afterwards.

  • Development of a MARS plugin for Jetbrains C# IDE “Rider”

    Content: MARS models need to be developed in the C# language with the .NET Core Runtime. We use “Rider” from JetBrains (creator of IntelliJ IDEA and such) for all our model development to have a unified development experience and greater compatibility across all major platforms. To further support and streamline the workflow of model developers, a MARS plugin for “Rider” has been developed. This plugin will connect certain parts of our microservice based infrastructure and allows to directly run new simulations from the IDE as well as debugging these. This will be a major improvement, since currently new models, will have to be uploaded through our WebUI and executed manually from it.
    Technologies: Jetbrains IDE API, Microservices, Docker
    Locality: Due to the tight integration requirements with our services, this thesis requires you to work in the MARS lab for at least 2-3 days a week. This is necessary to properly sync with your fellow students and get support from them 😉
    Kind: Bachelor Thesis, Master Thesis possible, if more features are added and a quick start is possible.

  • Integration of the contextual search engine MARS Surveyor into MARS

    Content: MARS Surveyor is a component, which allows to log events from a simulation and then perform a search based on contexts. The results are situations in the simulation, which are of a particular relevance to the modeller. This approach allows for a significantly more efficient way to explore large result sets from massive scale simulation models than simply browsing the data. Since MARS Surveyor has been developed in 2014 by Daniel Steiman (M.Sc.), the component needs to be refactored and extended to fit into todays MARS service landscape.
    Technologies: Context-Aware CEP, RabbitMQ, Docker
    Locality: Due to the integration requirements with our services, it is highly recommend to work in the MARS lab for at least 1-2 days a week. This is necessary to properly sync with your fellow students and get support from them 😉
    Kind: Bachelor Thesis, Master Thesis, if more in-depth research and a quick start are possible.

  • Optimization of the MARS Websuite User Experience

    • Consistent user interface design
    • Interaction- and workflow surveys
    • User questioning and usability lab
    • Usage of latest web technologies