On the development of Agent-Based Models for infrastructure evolution

Abstract

To achieve sustainability of our ever changing and evolving world system, we must view it as a control/engineering problem on a massive scale. Infrastructure systems are an essential part of this world and are critical to achieve sustain-ability. We conjecture that in order to sufficiently understand and manage these systems we need models that match their scale, complexity and evolutionary nature. In order to eventually capture the full diversity and adaptivity of the 'world system', and specifically the infrastructures that sustain it, we have developed an evolving, modular, approach for developing models of infrastructure evolution. This approach combines a staged and tested social process that enables model specification, a modular Java based simulation engine and an ontology, which is a formal language specification that uniquely defines the interface between minds (knowledge) and simulation models (executable code). The ultimate goal is to underpin decision-making in or on infrastructure systems. In the paper we will introduce the basic requirements necessary for creation of the evolving modeling and illustrate the adaptive approach for Agent Based Model development by showing a series of ever more complex models.