This paper addresses the question of the effects of the local modeling of agents and their decision-making on the behavior and the performance of these agents in a multiagent environment. Using the example of a particular architecture, the InteRRaP agent model, and a specific application, the loading-dock domain of interacting robots, a general model is presented that allows to predict the problem-solving and interaction performance of a specific type of InteRRaP agents, namely those controlled by local behavior-based decision functions. The analysis of the model provides evidence for the necessity of explicit mechanisms for coordination based on communication in many cases.
CITATION STYLE
Müller, J. P. (1996). A markovian model for interaction among behavior-based agents. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1037, pp. 376–391). Springer Verlag. https://doi.org/10.1007/3540608052_79
Mendeley helps you to discover research relevant for your work.