Architectures and applications of intelligent agents: A survey

Abstract

The objective of this paper is twofold. In its first part, we survey the state of the art in research on agent architectures. The architecture of an agent describes its modules and capabilities, and how these operate together. We structure the field by investigating three important research threads, i.e. architectures for reactive agents, deliberative agents and interacting agents. Then we describe various hybrid approaches that reconcile these three threads, aiming at a combination of different features like reactivity, deliberation and the ability to interact with other agents. These approaches are contrasted with architectural issues of recent agent-based work, including software agents, softbots, believable agents, as well as commercial agent-based systems. The second part of the paper addresses software engineers and system designers who are interested in applying agent technology to their problem domains. The objective of this part is to assist these readers in deciding which agent architecture to choose for a specific application. We characterise the most important domains to which the different approaches described in the first part have been applied, propose an application-related taxonomy of agents, and give a set of guidelines to select the right agent (architecture) for a given application.

An Autonomous Adaptive Multiagent Model for Building Automation

“…Self-regulating home appliances are considered reactive in the context of the AutoADMA model due to the reason that they have fixed pre-defined responses to certain sensor values. This layer is similar to Brooks' et al (1986) description of behavior modules, Ferguson' et al (1991) reactive layer and Müller' et al (1998) behavior-based component layer. This layer consists of a perception module, which extracts the sensory values from the sensors, an executor module that controls the actuators for action, and a reactive module that performs the function of control loops, which map the sensory values over the action commands and for that purpose use a reactive rule base.…”

“…It defines the decomposition of the overall system into a set of modules and their interaction, Kaebling et al (1991) whereby these modules and their interaction have to address, how sensor data and internal states of the agent determine the action. These architectures are characterized in Müller et al (1994) and Wooldridge et al (1995) as: behavior based architectures for instance Brooks et al (1986), Maes et al (1991) and Kaebling et al (1991), BDI based architectures, for instance Bratmen et al (1988) and Georgeff et al (1989) and hybrid architectures, for instance Ferguson et al (1991) and Müller et al (1998). Behavior-based architectures as described in Brooks et al (1986) are reactive to the environment through predefined response modules.…”

“…Within hybrid architectures, the behavior-based architectures are located at the lower layers for interaction with the environment and the BDI-based architecture are located at the higher layer for sophisticated reasoning and decision making. InterRAP, Müller et al (1998) and Touring Machine Ferguson et al (1991) are two widely known examples of hybrid architectures. These are quite similar in their layering scheme but different in their layered control mechanisms.…”

MAIA: An Event-Based Modular Architecture for Intelligent Agents

“…In the 1990s, research interest was focused on the investigation of architectural issues raised by three influential threads of agent research (i.e. reactive agents, deliberative agents and interacting agents), as collects the excellent survey by Müller [2].…”

Supervisor Localization: A Top-Down Approach to Distributed Control of Discrete-Event Systems

“…Two seminal studies -Brooks' subsumption architecture [8] and Arkin's motor schema [3] -introduced the prominent behavior-based paradigm, which in turn initiated a large number of investigations in intelligent agents [1,41] and cooperative robotics [2,10]. The behavior-based approach seems so promising that multi-robot teams have been increasingly applied to a variety of engineering domains.…”