InterJournal Complex Systems, 1810
Status: Accepted
Manuscript Number: [1810]
Submission Date: 2006
Modeling decentralized organizational change in honeybee societies
Author(s): Mark Hoogendoorn ,Martijn Schut ,Jan Treur

Subject(s): CX.6

Category:

Abstract:

The concept of organization has been studied in sciences such as Social Science and Economics, but recently also in Artificial Intelligence. With the desire to analyze and design more complex systems consisting of larger numbers of agents (e.g., in nature, society, or software), the need arises for a concept of higher abstraction than the concept agent. To this end, organizational modeling is becoming a practiced stage in the analysis and design of systems involving multiple agents. Hereby, the environment in which the multi-agent organization participates has to be taken into consideration. An environment can have a high degree of variability which might require organizations that change to adapt to the environment’s dynamics, to ensure a continuous proper functioning of the organization. Hence, such change processes are a crucial function of the organization and should be part of the organizational model. An organizational model incorporating organizational change can be specified in two ways: from a centralized perspective, in which there is a central authority that determines the changes to be performed within the organization, taking into account the current goals and environment. A second possibility is to create a model for organizational change from a decentralized perspective, in which each agent decides for himself if and how to change its own role allocations. In the latter approach, it is much more difficult for the organization as a whole to change in a coherent way, still satisfying the goals set for the organization, as there is no overall view of the organizational change. The approach might however be the only possibility for an organization to perform change as a central authority for performing change could be non existing or infeasible due to the nature of the organization. In the domain of social insects, such as honeybees and wasps, organizations are known to adapt in a decentralized fashion to environmental changes. This paper presents a generic model for decentralized organization change appropriate for such phenomena as occur in Nature. Such a model can aid in analyzing such an organization. The description of the model is done from a generic perspective, abstracting from the actual tasks being performed by the organization. The scope of the model is broader than being able to model adaptivity in societies of social insects: the mechanisms incorporated in the model facilitating decentralized organizational change may work in other types of organizations as well. Comparable approaches are for example used for finding an optimal allocation of cars to paint booths (see [1]). To evaluate the generic adaptive organization model being proposed, as a case study the honeybee (Apis Mellifera) has been investigated. For this domain the generic model has been instantiated. A number of different roles have been identified in the literature, of which five are addressed in the case study: (1) a brood carer takes care of feeding the larvae within the bee hive; (2) a patroller guards the hive by killing enemies entering the hive; (3) a forager which harvests food to be stored in the hive; (4) an undertaker cleans the hive of corpses, and (5) a resting worker simply does nothing. The mechanism used in the case study for switching between the different roles is a trigger mechanism by means of changes in the environment observed by the bees such as also identified in literature; e.g. [2]. In such a mechanism each role has a specific trigger, for which a bee has a certain threshold that determines whether this is the role it should play. The bee tends to plays the role for which it is most triggered. The instantiated model has been validated against properties as acquired from biological experts by means of checking these properties against simulation results of the instantiated model. [1] Bonebeau, E. and Theraulaz, G., Swarm Smarts, Scientific American, 282 (3): 72-79, 2000. [2] Theraulaz, G., Bonabeau, E., and Deneubourg, J.L., Response thresholds reinforcement and division of labor in insect societies. Proceedings of the Royal Society of London Series B-Biological Sciences, 265: 327-332, 1998.

Retrieve Manuscript
Submit referee report/comment


Public Comments: