An evolutionary behavioral model for decision making

For autonomous agents the problem of deciding what to do next becomes increasingly complex when acting in unpredictable and dynamic environments while pursuing multiple and possibly conflicting goals. One of the most relevant behavior-based models that tries to deal with this problem is the behavior network model proposed by Maes. This model proposes a set of behaviors as purposive perception–action units that are linked in a nonhierarchical network, and whose behavior selection process is orchestrated by spreading activation dynamics. In spite of being an adaptive model (in the sense of self-regulating its own behavior selection process), and despite the fact that several extensions have been proposed in order to improve the original model adaptability, there is not yet a robust model that can self-modify adaptively both the topological structure and the functional purpose of the network as a result of the interaction between the agent and its environment. Thus, this work proposes an innovative hybrid model driven by gene expression programming, which makes two main contributions: (1) given an initial set of meaningless and unconnected units, the evolutionary mechanism is able to build well-defined and robust behavior networks that are adapted and specialized to concrete internal agent’s needs and goals; and (2) the same evolutionary mechanism is able to assemble quite complex structures such as deliberative plans (which operate in the long-term) and problem-solving strategies. As a result, several properties of self-organization and adaptability emerged when the proposed model was tested in a robotic environment using a multi-agent platform.