A Nested Layered Threshold Model for Dynamic Task Allocation

Dynamic task allocation is an essential aspect in modeling ant behaviour [1],[2]. In this paper, we propose a new (abstract) model for dynamic task allocation of agents that combines a nested layered architecture with a threshold mechanism. We apply this new model to achieve flexible ant behaviour. To describe the actions of an agent, we describe a task allocation model based on behaviours and sub-behaviours that are associated with stimuli. At any time, an agent is performing one particular sub-behaviour within one behaviour. Changing stimuli may lead to reallocation of sub-behaviour or even the behaviour of the agent. A behaviour is a (course-grain) task an agent can perform. Behaviours can be seen as a collection of smaller, more primitive parts, called sub-behaviours.