About

We started this project as part of our Ph.D. work, because our research deals with multi-agent systems. Our research projects consider traffic and crowd simulation, and we try to simulate more more complex human-like behaviour.

Malte’s research depends on complex overtaking manoeuvre and norm usage on a high abstraction level. The research contains a more application-based structure, but it is a great and helpful work, in which way a multi-agent system can be used in complex simulation scenarios.
As an alumni of the DFG project SocialCars, his current academic research is focused on the field of (de-⁠)centralized cooperative traffic management and documented on the project’s GitHub-site.

Sophie’s research depends on social-choice theory which is the first usage of LightJason implementation. The generic aspect of this work is the aggregation of individual preferences into a collective decision.
As a member of the DFG project SocialCars, her current academic research is (like Malte’s) also embedded in the field of (de-⁠)centralized cooperative traffic management and documented on the project’s GitHub-site.

Philipp’s focus of research is the high scalability of such complex simulation systems with data mining / big data and software developing aspects. His thesis mainly contains algorithms, data structure, software developing and formal structure. The applied software building process is inspired by Scrum and Continuous Integration to get a long-live, robust software architecture.

Based on these perspectives, we can motivate our work. Our work is inspired by Jason, a multi-agent framework developed by Jomi F. Hübner and Rafael H. Bordini, wich we didn’t use due to conflicts with our requirements. Instead we based our work on some of their theoretical concepts and ideas.

We tried a lot of different types of frameworks like TerraME, FLAME, GOAL or MASON to create an usable system for our simulation, but without success, so that we build up our framework from scratch.

From another point of view, we had compared the main popular programming language indices Tiobe, RedMonk and PopularitY in 2015 / 2016 and we found out, that agent-programming languages are not listed on any index, only logical programming (Prolog) is found on Tiobe. Our goal for this system is to create a practically usable and well documented framework which can also be used by non-software-developers and included in already existing projects. We try to explain the ideas, concepts and technical background in a clean and understandable way to get a better approval of agent-programming and modelling.

Except where otherwise noted,

Efficient Beliefbase previous page
next page contact