Tristan M. Behrens

Towards an environment interface standard for agent platforms

We introduce an interface for connecting agent platforms to environments. This interface provides generic functionality for executing actions and for perceiving changes in an agent’s environment. It also provides support for managing an environment, e.g., for starting, pausing and terminating it. Among the benefits of such an interface are (1) standard functionality is provided by the interface implementation itself, and (2) agent platforms that support the interface can connect to any environment that implements the interface. This significantly reduces effort required from agent and environment programmers as the environment code needed to implement the interface needs to be written only once. We propose that the interface presented may be used as a standard that enables agents to control entities in environments. Our starting point for designing such a generic interface is based on a careful study of the various interfaces used by different agent programming languages to connect agent programs to environments. We discuss several case studies that use our interface (an elevator simulator, the well-known agent contest, and an implementation of the interface to connect agents to bots in Unreal Tournament 2004).

An interface for agent-environment interaction

Agents act and perceive in shared environments where they are situated. Although there are many environments for agents --- ranging from testbeds to commercial applications --- such environments have not been widely used because of the difficulty of interfacing agents with those environments. A more generic approach for connecting agents to environments would be beneficial for several reasons. It would facilitate reuse, comparison, the development of truly heterogeneous agent systems, and increase our understanding of the issues involved in the design of agent-environment interaction. To this end, we have designed and developed a generic environment interface standard. Our design has been guided by existing agent programming platforms. These platforms are not only suitable for developing agents but also already provide some support for connecting agents to environments. The interface standard itself is generic, however, and does not commit to any specific platform features. The interface proposal has been implemented and evaluated in a number of agent platforms. We aim at a de facto standard that might become an actual standard in the near future.

Unreal Goal Bots

It remains a challenge with current state of the art technology to use BDI agents to control real-time, dynamic and complex environments. We report on our effort to connect the GOAL agent programming language to the real-time game UNREAL TOURNAMENT 2004. We focus in particular on the design of a suitable interface to manage agent-bot interaction and argue that the use of a recent toolkit for developing an agent-environment interface provides many advantages.It remains a challenge with current state of the art technology to use BDI agents to control real-time, dynamic and complex environments. We report on our effort to connect the GOAL agent programming language to the real-time game UNREAL TOURNAMENT 2004. BDI agents provide an interesting alternative to control bots in a game such as UNREAL TOURNAMENT to more reactive styles of controlling such bots. Establishing an interface between a language such as GOAL and UNREAL TOURNAMENT, however, poses many challenges. We focus in particular on the design of a suitable and reusable interface to manage agent-bot interaction and argue that the use of a recent toolkit for developing an agent-environment interface provides many advantages. We discuss various issues related to the abstraction level that fits an interface that connects high-level, logic-based BDI agents to a real-time environment, taking into account some of the performance issues.

The Multi-Agent Programming Contest

The international Multi-Agent Programming Contest (MAPC), is a community-serving effort to facilitate advances in programming multiagent systems (MAS) by (1) developing benchmark problems, (2) enabling head-to-head comparison of MAS’s and (3) supporting educational efforts in the design and implementation of MAS’s. The tournament platform, MASSim, is freely available and we encourage everybody to download it and use it in the classroom

The multi-agent programming contest 2011: a résumé

The Multi-Agent Programming Contest is an annual AI competition aiming at comparing deliberative techniques for problem solving, that are based on formal approaches and computational logics. In 2011 the Contest was held for the seventh time and witnessed the introduction of the new agents on Mars scenario. We give an overview of the Contest in general but concentrate on the agents on Mars scenario. We also provide empirical results that we received before, during and after the tournament.

The multi-agent programming contest from 2005---2010

The Multi-Agent Programming Contest is an annual international event on programming multi-agent systems: Teams of agents participate in a simulated cooperative scenario. It started in 2005 and is organised in 2010 for the sixth time. The contest is an attempt to stimulate research in the area of multi-agent system development and programming by (i) identifying key problems in the field and (ii) collecting suitable benchmarks that can serve as milestones for testing multi-agent programming languages, platforms and tools. This article provides a short history of the contest since it started and reports in more detail on the cows and cowboys scenario implemented for the 2008, 2009 and 2010 contest editions. We briefly discuss the underlying technological background and conclude with a critical discussion of the experiences and lessons learned.

Query Caching in Agent Programming Languages

Agent programs are increasingly widely used for large scale, time critical applications. In developing such applications, the performance of the agent platform is a key concern. Many logic-based BDI-based agent programming languages rely on inferencing over some underlying knowledge representation. While this allows the development of flexible, declarative programs, repeated inferencing triggered by queries to the agent’s knowledge representation can result in poor performance. In this paper we present an approach to query caching for agent programming languages. Our approach is motivated by the observation that agents repeatedly perform queries against a database of beliefs and goals to select possible courses of action. Caching the results of previous queries (memoization) is therefore likely to be beneficial. We develop an abstract model of the performance of a logic-based BDI agent programming language. Using our model together with traces from typical agent programs, we quantify the possible performance improvements that can be achieved by memoization. Our results suggest that memoization has the potential to significantly increase the performance of logic-based agent platforms.

Agent-oriented control in real-time computer games

We introduce a new methodology that allows to apply techniques and methods from agent-oriented programming (AOP) to computer games. To this end, we introduce an underlying model based on agents, controllable entities, and an agents-entities-relation. Then we elaborate on a method how to steer several controllable entities through an environment with a single agent. Finally, we show how AOP can be used to implement several agents based on the method.

Multi-agent navigation using path-based vector fields

We present an approach to multi-agent navigation, that is based on generating potential-fields from A*-paths. We will introduce and compare two algorithms: 1) a geometrical algorithm that is based on quads, and 2) an images-based algorithm. We show empirically that the images-based algorithm is more memory-consuming, but has better performance.