Lars Braubach

Jadex: A BDI Reasoning Engine

This chapter presents Jadex, a software framework for the creation of goal-oriented agents following the belief-desire-intention (BDI) model. The Jadex project aims to make the development of agent based systems as easy as possible without sacrificing the expressive power of the agent paradigm. The objective is to build up a rational agent layer that sits on top of a middleware agent infrastructure and allows for intelligent agent construction using sound software engineering foundations. Fostering a smooth transition from traditional distributed systems to the development of multi-agent systems, well established object-oriented concepts and technologies such as Java and XML are employed wherever applicable. Moreover, the Jadex reasoning engine tries to overcome traditional limitations of BDI systems by introducing explicit goals. This allows goal deliberation mechanisms being realized and additionally facilitates application development by making results from goal-oriented analysis and design easily transferable to the implementation layer. The system is freely available under LGPL license and provides extensive documentation as well as illustrative example applications.

Goal representation for BDI agent systems

Agent-oriented system development aims to simplify the construction of complex systems by introducing a natural abstraction layer on top of the object-oriented paradigm composed of autonomous interacting actors. One main advantage of the agent metaphor is that an agent can be described similar to the characteristics of the human mind consisting of several interrelated concepts which constitute the internal agent structure. General consensus exists that the Belief-Desire-Intention (BDI) model is well suited for describing an agents mental state. The desires (goals) of an agent represent its motivational stance and are the main source for the agents actions. Therefore, the representation and handling of goals play a central role in goal-oriented requirements analysis and modelling techniques. Nevertheless, currently available BDI agent platforms mostly abstract from goals and do not represent them explicitly. This leads to a gap between design and implementation with respect to the available concepts. In this paper a generic representation of goal types, properties, and lifecycles is developed in consideration of existing goal-oriented requirements engineering and modelling techniques. The objective of this proposal is to bridge the gap between agent specification and implementation of goals and is backed by experiences gained from developing a generic agent framework.

Jadex: A BDI-Agent System Combining Middleware and Reasoning

Nowadays a multitude of different agent platforms exist that aim to support the software engineer in developing multi-agent systems. Nevertheless, most of these platforms concentrate on specific objectives and therefore cannot address all important aspects of agent technology equally well. A broad distinction in this field can be made between middleware- and reasoning-oriented systems. The first category is mostly concerned with FIPA-related issues like interoperability, security and maintainability whereas the latter one emphasizes rationality and goal-directedness. In this paper the Jadex reasoning engine is presented, which supports cognitive agents by exploiting the BDI model and is realized as adaptable extension for agent middleware such as the widely used JADE platform.

A goal deliberation strategy for BDI agent systems

One aspect of rational behavior is that agents can pursue multiple goals in parallel. Current BDI theory and systems do not provide a theoretical or architectural framework for deciding how goals interact and how an agent can decide which goals to pursue. Instead, they assume for simplicity reasons that agents always pursue consistent goal sets. By omitting this important aspect of rationality, the problem of goal deliberation is shifted from the architecture to the agent programming level and needs to be handled by the agent developer in an error-prone ad-hoc manner. In this paper a goal deliberation strategy called Easy Deliberation is proposed allowing agent developers to specify the relationships between goals in an easy and intuitive manner. It is based on established concepts from goal modeling as can be found in agent methodologies like Tropos and requirements engineering techniques like KAOS. The Easy Deliberation strategy has been realized within the Jadex BDI reasoning engine and is further explained by an example application. To fortify the practical usefulness of the approach it is experimentally shown that the computational cost for deliberation is acceptable and only increases polynomially with the number of concurrent goals.

Extending the capability concept for flexible BDI agent modularization

Multi-agent systems are a natural way of decomposing complex systems into more manageable and decentralized units. Nevertheless, as single agents can represent complex subsystems themselves, software engineering principles for the design and implementation of coherent parts of single agents are necessary for producing modular and reusable software artifacts. This paper picks up the formerly proposed capability concept for structuring BDI agents in functional clusters, and generalizes and extends it to support a higher degree of reusability. The resulting mechanism allows for designing and implementing BDI agents as a composition of configurable agent modules (capabilities). It is based on a black-box approach with export interfaces that is in line with object-oriented engineering principles.

A flexible BDI architecture supporting extensibility

The BDI agent model comprises a simple but efficient folk psychological framework of mentalistic notions usable for modeling rational agent behaviour. Nevertheless, despite its usefulness it is also a popular subject for extensions that try to improve the model in certain uncovered aspects such as emotions or norms. On the architectural level the BDI model is typically represented by an abstract BDI interpreter, which implements the fixed BDI reasoning cycle. In this paper it is argued that a fixed cycle has certain inherent drawbacks and that a transition towards a flexible agenda approach based on BDI meta-actions leads to a design open for extensions in many respects because new meta-actions can be easily integrated into the architecture on demand. To prove the validity of the approach, it is shown how the extensibility can be exploited to integrate concrete new aspects of increasing complexity into the model. They range from a simple mechanism for updating beliefs to a complex goal deliberation strategy and demand only slight modifications at well-defined extension points of the architecture. The new architecture as well as the presented extensions has been realized within the open source Jadex BDI reasoning engine.

Evaluation of agent–oriented software methodologies – examination of the gap between modeling and platform

More and more effort is made to provide methodologies for the development of agent-based systems. Awareness has grown that these are necessary to develop high quality agent systems. In recent years a number of proposals have been given. Based on our experiences we argue that a complete evaluation of methodologies cannot be done without considering target platforms, because the differences between available implementations are too fundamental to be ignored. In order to conduct a suitable comparison we present a flexible evaluation framework that takes platform specific criteria into account. Part of this framework is a procedure to derive relevant criteria from the evaluated platforms and methodologies. In combination with a set of platform dependent and independent criteria our framework allows evaluation of the appropriateness of methodologies with respect to platforms. As a consequence, also the suitability of methodologies for an individual platform, or vice versa of several platforms for an individual methodology can be examined. To show the usefulness of our proposal, we evaluate the suitability of different methodologies for an example platform.

Validation of BDI agents

Testing and Debugging multi-agent systems (MAS) - which are inherently concurrent and distributed - is a challenging task. While complex application scenarios demand intelligent entities with autonomous reasoning capabilities, the applied reasoning mechanisms impair current approaches to validate MAS implementations. Reactive planning systems, namely the well-known Belief Desire Intention (BDI) architecture, have been successfully applied to implement these intelligent entities by means of goal directed agents. Despite testing and debugging, used to validate whether implementations behave as intended, are crucial to serious development efforts, only minor attention has been payed to corresponding tool support and testing procedures for BDI-based MAS. In this paper, we examine how the reasoning mechanism inside agent implementations can be checked and how static analysis of agent declarations can be used to visualize and check the overall communication structure in closed MAS. We present corresponding tool support, which relies on the definition of crosscutting concerns in BDI agents and enables both approaches to the Jadex Agent Platform.

Augmenting BDI agents with deliberative planning techniques

Belief-Desire-Intention (BDI) agents are well suited for autonomous applications in dynamic environments. Their precompiled plan schemata contain the procedural knowledge of an agent and contribute to the performance. The agents generally are constrained to a fixed set of action patterns. Their choice depends on current goals, not on the future of the environment. Planning techniques can provide dynamic plans regarding the predicted state of the environment. We augment a BDI framework with a state-based planner for operational planning in domains where BDI is not well applicable. For this purpose, the requirements for the planner and for the coupling with a BDI system are investigated. An approach is introduced where a BDI system takes responsibility for plan monitoring and re-planning and the planner for the creation of plans. A fast state-based planner utilizing domain specific control knowledge retains the responsiveness of the system. In order to facilitate integration with BDI systems programmed in object-oriented languages, the planning problem is adopted into the BDI conceptual space with object-based domain models. The application of the hybrid system is illustrated using a propositional puzzle and a multi agent coordination scenario.

The active components approach for distributed systems development

The development of distributed systems is an intricate task due to inherent characteristics of such systems. In this paper these characteristics are categorised into software engineering, concurrency, distribution and non-functional criteria. Popular classes of distributed systems are classified with respect to these challenges, and it is deduced that modern technological trends lead to the inception of new application classes with increased demands regarding challenges from more than one area. One recent example is the class of ubiquitous computing, which assumes dynamic scenarios in which devices come and go at any time. Furthermore, it is analysed to which extent todays prevailing software development paradigms – object, component, service and agent orientation – are conceptually capable of supporting the challenges. This comparison reveals that each of the paradigms has its own strengths and weaknesses and none addresses all of the challenges. The new active component approach is proposed aiming at a conceptual integration of the existing paradigms in order to tackle all challenges in an intuitive and unified way. The structure, behaviour and composition of active components are explained, and an infrastructure for active components is introduced. To underline the usefulness of the approach real-world applications is presented and an evaluation according to the challenges is given.