Cellular automata models for cooperation in multirobot systems
IMMURO'12 Proceedings of the 11th WSEAS international conference on Instrumentation, Measurement, Circuits and Systems, and Proceedings of the 12th WSEAS international conference on Robotics, Control and Manufacturing Technology, and Proceedings of the 12th WSEAS international conference on Multimedia Systems & Signal Processing
| Hi-index | 0.00 |
This dissertation studies the problems of modeling and analysis of the behavior of large-scale multi-agent systems. We model a broad variety of multi-agent systems as appropriate variants of cellular and network automata. We then analyze various aspects of the collective dynamics of those cellular and network automata. In that endeavor, we take advantage of motivations, paradigms, models and analytical tools ranging from theoretical computer science and formal methods to distributed artificial intelligence to agent-based modeling to complex dynamical systems.More specifically, we focus on three important aspects of large-scale distributed information systems. First, we address the temporal and causal nature of inter-agent interactions, and its implications. In that context, we compare and contrast cellular automata with different communication models. Second, we study the implications of homogeneity vs. heterogeneity of the individual agent behaviors in a large multi-agent system. Third, in conjunction with the models of the individual agent behaviors, we also analyze the impact of the communication network topology on the collective dynamics of an ensemble of autonomous agents. In particular, we establish that a number of fundamental problems about multi-agent systems, abstracted as appropriate types of network automata, are demonstrably computationally intractable, even when instances of the network automata models under scrutiny are severely constrained.Our work strengthens and/or generalizes a number of results on the global behavior of various dynamical system models studied in the literature, such as the classical cellular automata and discrete Hopfield networks. Among several far-reaching implications of our results, we emphasize the general conclusion that a highly complex and generally unpredictable collective behavior can arise from the synergy of very simple individual behaviors and their loosely coupled local interactions. Keywords. multi-agent systems, distributed artificial intelligence, theoretical computer science, formal methods, computational complexity, cellular and network automata, agent-based modeling, discrete dynamical systems, complex systems