Malte Aschermann’s research while affiliated with Technische Universität Clausthal and other places

Current BDI agent frameworks often lack necessary modularity, scalability and are hard to integrate with non-agent applications. This paper reports ongoing research on LightJason, a multi-agent BDI framework based on AgentSpeak(L), fine-tuned to concurrent plan execution in a distributed framework; LightJason aims at efficient and scalable integration with existing platforms. We state requirements for BDI agent languages and corresponding runtime systems, and present the key concepts and initial implementation of LightJason in the light of these requirements. Based on a set of requirements derived for scalable, modular BDI frameworks, the core contribution of this paper is the definition of a formal modular grammar for AgentSpeak(L++), a modular extension of AgentSpeak(L), and its underlying scalable runtime system. A preliminary validation of LightJason is given by means of an example evacuation scenario, an experimental analysis of the runtime performance, and a qualitative comparison with the Jason platform.

In 2013, 2+1 roadways have become mandatory for newly constructed rural roadways in Germany. The steady trend towards autonomous vehicles and vehicle-to-X (V2X) communication will enable new automated traffic coordination mechanisms. In our research, we study how traffic flow on 2+1 roadways can be improved by using such mechanisms for coordinating the usage of overtaking lanes. Conflicts between vehicles on 2+1 roadways arise due to (1) differing capabilities (e.g. maximum speed), and (2) conflicting preferences of drivers (e.g. desired speed). These conflicts can lead to poor use of resources, and ultimately in time loss both from an individual vehicle and aggregated system perspective. To foster acceptance of coordination mechanisms, it is important to not only consider the system perspective but also to ensure some level of driver satisfaction and fairness, e.g. if vehicles are denied access to the overtaking lane in favour of faster vehicles, this raises questions regarding the acceptance of external coordination. While there is some research studying microscopic optimisation of traffic flows on rural 2+1 roads and intersections, there has been only little research focusing on driver satisfaction and fairness of mechanisms. In this work, we present the results of a pre-study on simulated 2+1 manoeuvres conducted using the Simulation of Urban Mobility (SUMO) suite to estimate optimisation potentials of coordination. Preliminary results indicate potentials to reduce driver dissatisfaction while maintaining fairness. We analyse optimisation potentials and propose a model which combines driver satisfaction and fairness of coordination, aiming at increasing acceptance of autonomous vehicle coordination.

In this paper we address the problem of retrieving similar resources which are distributed over a multi-agent system (MAS). In distributed environments identification of resources is realized by using cryptographic hash functions like SHA-1. The issue with these functions in connection with similarity search is that they distribute their hash values uniformly over the codomain. Therefore such IDs cannot be used to estimate the similarity of resources, unless one enumerates the whole search space and retrieves every resource for comparison. In this paper we present a three-layer architecture and a data model to efficiently locate similar resources in linear time complexity by using locality-sensitive hash functions. We design the data model as an extension to distributed environments (MAS), which only need to provide at least basic resource management capabilities, such as storing and retrieving resources by their ID. We use a benchmark data set to compare our approach with state-of-the-art centralized heuristic approaches and show that, while these approaches provide better search accuracy, our approach can deal with decentralized data and thus, allows us to flexibly adapt to dynamic changes in the underlying MAS by distributing and updating sets of information about similarities over different agents.