An aspect-oriented approach for modeling self-organizing emergent structures

Abstract

Multi-agent systems must be engineered to ensure that desirable system-level properties will consistently emerge from the complex interactions of the underlying agents, while also guaranteeing that undesirable behavior will be suppressed. We present an Aspect-Oriented Programming (AOP) framework for modeling, visualizing and manipulating emergent structure in multi-agent systems. By encapsulating the macroscopic structure, we can identify undesirable patterns of behavior at a higher level of abstraction. The identification of such patterns allows us to implement a feedback loop to steer the behavior of the lower level agents towards actions favorable for the emergence of a reliable solution.AOP facilitates the modeling of the system-wide behavior, thus it serves as a valuable tool for building confidence that a given multi-agent system will consistently meet its requirements.

Weaving the Fabric of the Control Loop through Aspects

“…Another set of approaches proposes an AOP language as the primary mechanisms for expressing monitoring, analyis and adaptation policies [32,31,30,6,13,16,4]. Shomrat et al [32] and Serban et al [31] show how architectural principles can be automatically enforced through first-class aspects.…”

“…Shomrat et al [32] and Serban et al [31] show how architectural principles can be automatically enforced through first-class aspects. Seiter et al [30] proposes to use aspects for modeling and controlling the macroscopic system-wide behavior of a multi-agent system. David et al [6] proposes a domain-specific, event-based AOP language for expressing event-condition-action policies as first-class aspects and use the Fractal component model for implementing the adaptations.…”

A cybernetic approach to multi-agent system simulation in Tijuana-San Diego using the Viable Systems Model

“…Increasing emphasis is being paid to macro-level constructs -that is accounting for the context in which agents are operating, although this work is currently constrained by a lack of well defined models of context Seiter et al, [10]. This inability to appreciate social systems at the macro level presents a major limitation to any use of MAS systems to understand social problems in the Tijuana-San Diego metropolis.…”

“…Seiter et al, [10] affirm this trend in arguing that agent-oriented methodologies to date have primarily focused on techniques for engineering agents such as coding of their beliefs, goals, rules and interactions, with less focus of engineering the emergent macroscopic behaviour that arises from the agents interactions, related to complex properties such as autonomy, learning, adaptation, mobility, interaction and collaboration. This they argue is primarily due to the lack of well-defined models and processes for dealing with the macro-level.…”

“…In terms of the benefits that consideration of VSM constructs could bring to research on MAS, a useful approach with micro-level and macro-level considerations is that already alluded to by Seiter et al [10]. In terms of micro-level considerations, these would constitute evaluation of any concept, structure or function of the VSM at any particular layer of recursion for a particular multi-agent system -given that the VSM comprises systems within systems.…”