Michael Papasimeon

Extending the UML for designing Jack agents

Mainstreaming and industrialising agent technologies requires suitable methodological and technological support for the various engineering activities associated with managing the complexity of any software system development. Despite its origins in object oriented software engineering the UML provides a rich and extensible set of modelling constructs that can be applied to agent oriented technologies. This paper provides details of extensions to the UML for the design of agents that are to be implemented in the JACK language. These extensions provide the capacity to model the behaviour of agents for the purposes of design and, though the extensions are language specific, future generalisation and application to other agent languages can be supported as a industry-wide consensus about the nature of agency emerges over the next few years. This research builds on previously proposed extensions to the UML and moves a step closer to the goal of providing through-life engineering support to agent oriented systems development. This work is motivated by a pressing need to maintain, modify, develop and deploy existing and future agent based simulations of military operations for the Australian Defence Force.

Specifying Agent Behaviour with Use Cases

The software engineering of multi-agent systems demands specification of the required agent behaviours to provide documented requirements for the design and implementation phases. A methodology for the analysis and specification of agent behaviours is proposed, which arises from a lengthy experience in the construction of multi-agent simulations for military operations research. The methodology builds upon the existing use case modelling techniques provided by the Unified Modeling Language (UML) and is in keeping with the agent extensions to the UML proposed elsewhere. A case-study from a specific multi-agent air combat simulation accompanies the elaboration of the methodology.

The human agent virtual environment

In this paper we describe a multi-agent simulation called the Human Agent Virtual Environment (or HAVE). HAVE is a test bed to explore agent-environment interaction in multi-agent simulation for defence applications. The primary research driver in the development of HAVE is to explore representations of virtual environments in which both humans and agents are situated, perceive these environments and undertake meaningful and appropriate actions. HAVE models and simulates a Close Air Support (CAS) mission which involves fighter or strike aircraft providing support to ground troops through the use of air-to-ground weapons. This provides a realistic and currently extremely relevant domain in which to explore agent-environment interactions. Three important research challenges have been addressed by the work. The first, is the implementation of a multi-modal representation of the virtual environment, having multiple, parallel yet consistent representations of the virtual world that were accessible to, and tailored for the different simulation components. The second is the used of labeled annotations in the virtual world which the agents could easily access and interpret. The third, the use of an appropriate architecture for capturing and representing interaction between agents and the environment they are situated in.

Simulating Fighter Pilots

Since 1990 a focused intelligent agent research and development programme within the Defence Science and Technology Organisation (DSTO) has underpinned a strong history of deployed operational simulations. Originally aimed at improving simulations of fighter pilots the research has expanded to include: fundamentals of agent languages and architectures; the cognition of teams; intention recognition and cognitive modelling; simulating civilian behaviour in conflict; intelligent environments; software engineering; and autonomy and uninhabited aerial vehicles. Capitalising on this research are a series of deployed simulations that have provided strong examples of the benefits of the technology. This paper presents a brief account of four successful agent-based simulation systems and a broad but shallow overview of some of the more interesting aspects of our relevant agent research and development activities.

Real--Time UAV Maneuvering via Automated Planning in Simulations

The automatic generation of realistic behaviour such as tactical intercepts for Unmanned Aerial Vehicles (UAV) in air combat is a challenging problem. State-of-the-art solutions propose hand–crafted algorithms and heuristics whose performance depends heavily on the initial conditions and specific aerodynamic characteristics of the UAVs involved. This demo shows the ability of domain–independent planners, embedded into simulators, to generate on–line, feed–forward, control signals that steer simulated aircraft as best suits the situation

Issues in modelling sensor and data fusion in agent based simulation of air operations

Sensorfusion is often not the focus of opera- tional analysis, but the increased connectivity and infonna- tion sharing of a network enabled wagare environment and plagonns with sophisticated on-board sensor fusion capa- bilities makes it an important consideration for military an- alysts. Modelling sensor fusion within a simulation of large- scale military operations is complicated because the sen- sor fusion processes under consideration are not conjned to the mission systems of the various platforms. The cog- nitive processing of sensor data by air-crews and the entire command and control chain are as important. Agent based techniques can offer architectural modelling options that, as well as providing flexibility in meeting the requirements of simulation developers, can address issues raised by the wider sensor fusion communig. This paper broadens accepted definitions of sensor fu- sion to characterise it as a socio-technical problem requir- ing that attention be given to the human users of sensor fusion systems as well as the technical aspects internal to the mission systems. A set of models thar illustrates this approach is presented. This modelling approach supports the adoption of intelligent agent techniques and technolo- gies. Intelligent agent architectures for modelling sensor fusion together with insights into particular implementa- tions that show promise in addressing the needs of the sim- ulation community and may offer insights for developers of sensor fusion systems.

Towards reuse in agent oriented information systems: the importance of being purposive

The emergence of large information systems has pushed software specification into the area of business modelling to adequately capture and consider business requirements. At the same time, there has been a move toward techniques for specifying the behaviours of and the knowledge associated with intelligent agents as these are increasingly found as important components of those information systems. This paper presents four software models useful for specifying the requirements of an agent oriented information system. Adopting a similar notation for each model smooths the transition between models. It will be shown that it is in the relationships between these models there is scope for capturing purposive descriptions that facilitate reuse at various levels. A commentary on the importance of an explicit representation of the purpose for which a software component is intended is provided followed by an example, from the domain of military simulation, that illustrates the model and its application. The aim of this paper is to present a modelling approach that unifies business models, use case models, agent behavioural models and domain models, for the purpose of specifying an agent oriented information system.

Discovering Emergent Agent Behaviour with Evolutionary Finite State Machines

In this paper we introduce a novel approach to discovering emergent behaviour in multiagent simulations, using evolutionary finite state machines to model intelligent agents in an adversarial two-player game. Agent behaviour is modelled as a finite set of predetermined states. The logic that leads to transitions between states is evolved to maximise fitness, which is determined through execution in a constructive simulation environment. The resultant evolved finite state machine (E-FSM) is evaluated for two finite state machine implementations, one with states specifically designed to perform a known behaviour and the other with states consisting of generic actions. Our experiments demonstrate that this approach can discover complex emergent behaviours from simple, generic actions, and use these behaviours to achieve a position of tactical superiority in the domain of air combat simulation.

Augmented Reality for Multi-agent Simulation of Air Operations

This paper demonstrates the potential of augmented reality to enhance visualisation and facilitate interrogation of multi-agent simulations of military operations. It describes four use cases for augmented reality in the context of constructive and virtual simulation for operations research, and discusses work undertaken and proposed future projects in both small and large augmented environments. The prototypes developed are used to assess the feasibility and benefits of applying augmented reality to simulation visualisation, wargaming, and the explainability and interpretability of agent behaviour.

Software Engineering Multi-Agent Simulations for Computational Operations Research: Challenges and Lessons from 20 Years of Practice

This paper presents challenges faced and lessons learnt from over twenty years experience with the software engineering of multi-agent simulations for computational operations research of air to air combat in the Defence Science and Technology Organisation (DSTO). A major role of the DSTO is to undertake research and development to support Australias current and future air combat capability.