Game Engines and MAS: Tuplespace-based Interaction in Unity

   page       attach   

Game Engines are gaining more and more importance in both industrial field - enabling an easy development of modern applications and videogames - and in the research field, with a particular regard to multi-agent systems (MAS). Their expressive power, along with modern technologies support and ad- vanced functionalities, allow the design and creation of complex systems ef- ficiently. While becoming more stable and supporting new functionalities, Game Engines are nowadays a well-suited reality, therefore reliable in order to create modern, advanced systems and applications (e.g. augmented/virtu- al/mixed reality, immersive simulations, modern videogames, 3D world de- sign).
Nevertheless, they still lack of proper general abstractions, which could be suitable to be used for tackle complexity when designing and implementing complex systems.
For this purpose, the attempt to exploit Game Engines as integrated de- velopment environments where MAS abstractions are well-suited to tackle complexity is worth to be done, in order to bring agent-oriented software engineering and coordination models as main providers of new solutions and solving technology problems with the aid of Game Engines.
In this dissertation, we analyse the Unity3D Game Engine and present two prototype API, meant to exploit a previous integration of Prolog within the same framework to provide Unity3D of a new abstraction level, enabling a tuplespace based interaction and coordination model to be used by program- mers via LINDA primitives.
Those libraries are organized around Unity3D features, providing an easy- to-use and accessible way to approach coordination and interactions among objects when building complex systems with Unity3D, bringing all those properties and advantages that are well-established in MAS context, both from the complex systems engineering point of view and management of so- cial interactions side.
All new functionalities are tested using proper case studies, which illustrate their expressiveness, correctness and efficacy.

keywordsmulti-agent system, game engines, Unity3D, Prolog, tuple space, coordination, Linda, interaction, agent society, MAS model