Code: (11.y.COD)


Title: ProtozOOP Coding


/// Authors: Sergio del Castillo + Eva Castiñeira + JARD

/// Year: 2011

/// Prmtrs: Textual Aural   Translation     Immanence       Days    Collaborative   Participatory   Exogenous        Massive Exposed


Each agent has an associated ID and they are programmed with a ROUTINE (behavior), an ALERT (satisfaction or dissatisfaction depending on its behavior) and a DOMAIN (entities directly affected by its behavior). Installation users are identified with every agent and they have to manage the negotiation to reach the ecosystem equilibrium so that the installation works out like an assistant for reaching agreements.




Methodologies (Protocol).   


To program a Virtual Ecosystem it is necessary to elaborate a MAS diagram. A multi-agent system (MAS) is a distributed system where the combined behavior of those elements or agents produces a result “intelligent” as a whole, able to tend to achieve any objective. It is important to understand that agents (ID’s) are not necessarily intelligent. There are two types of multi-agent systems generations, the formal one and the constructivist one; the latter, which is our case, tries to provide all the agents as a whole with intelligence. This way, through interactive elaborated mechanisms (programmed routines), the system itself generates intelligent behavior which was not necessarily planned at the beginning or defined in the own agents (which can be really simple). This kind of global behavior is habitually called emergent behavior.


      a.1. Descomposition. The MAS is characterized by Autonomy: agents are at least partially autonomous, each one has a goal to achieve; by Local vision: agents have no global vision of the system, or the system is too complex for an agent to make some practical use of this knowledge; Decentralization: there is not any assigned system global agent, all of them contribute to the whole, there can only be prefixed restrictions by aprioristic consensus.

      a.2. Formulation. An example of agent formulation is the agent ID00; Routine: to examine front views through its rule. Alert: When it finds in front an obstruction (another entity which is within its domain). Domain: agents 1, 2, 4, 9 (agents taken as possible computable obstructions).  

      a.3. Modelling. Based on the agent adaptive informality (as well named AIA, Abstract Intelligent Agents), the represented iconic forms associated with every agent vary according to the routines used for programming them, showing at every moment and continuously the system fluctuation, they are not immovable preexistences and always they depend as results of the whole as far as their nature allow us.

      a.4. Integration: assembly and approximation. All and every one of the agents can be understood as pulsions of each one of the disciplines involved in the project, or understood as clusters or subsystems or analyzable parts separately or cluster parts of a concrete project (closing system agent, structure agent, climate agent, place or surrounding conditions agent, etc.) The multi-agent system allows them to come to an agreement, in contact, in conversation, while none of them dominates the others without having understood them as a whole. The choice of the order and importance of agents to form the project is part of the project itself, and it becomes more transparent as decision taken as rule achieve in a consensus, if it existed.       

      a.5. Exploration. The user manipulates, through augmented reality markers (in our case, Reactivision fiduciary markers, open source) controlled by webcam, the entities (ID’s) position, whose grouping forms a protozOOP, which is this topological scheme based on relations and transactions among users. This way, users (who can embody those disciplines, departments, involved agents) get involved as protozOOPs programmer within this visual laboratory.                                                    

      a.5.a-Register of objectives and stopping conditions: Every phase could mean to discover a synergy among agents not considered previously, which can be interested to be conserved; those achievements are fixed, or those mistakes are corrected, while the rest of agents in the system keep looking to reach their partial objectives, but starting from this new scenario

      a.5.b- Cycle and recalculation conditions: there is a continuous computation in closed-cycle of the recalculation of relations and analysis and action until a stopping condition is reached, as for instance, reaching an optimum behavior among the agents in at least a 90% of cases.      


Article and Publication/ Colaboratorio Book CoL:




Fan VK: