Recently, I have done a work about Agent-Based Computational Economics and the Object-Field model as a novel way to explore Economic Systems. This is my (with Haris) proposal:
Economies are complex adaptive systems encapsulating micro structures and behaviors, interaction patterns, and macroscopic regularities. Thus, studies of economic systems must consider how to handle interdependent feedback interactions of micro behaviors, interaction patterns and macroscopic regularities.
One such approach is the Agent-Field approach of the agent-based computational economics. In this framework, models of economic systems are viewed as collections of multi-scale and structured economic agents from the real world such as individuals, social groupings, institutions and physical entities and as smooth, continuous economic environments called fields. In other words, the Agent-Field framework is a fused agent-based model by capturing agents in indeterminate economic environments conceptualized as continuous, differentiable fields with variable levels of spatial uncertainty and embedded semantics. The science of the Agent-Field model is drawn from the field of Geographic Information Science (GIS) models (particularly the Object-Field model) and the field of Agent-Based Computational Economics. Thus, a common base-model for the Agent-Field framework is proposed by giving it a formal definition using the Unified Modeling Language (UML). We explore potential advantages and disadvantages of the Agent-Field framework for the study of economic systems using the EU leasing market economy as an example of demonstrating the application of the framework. This also enables us to formulate an initial frame representation of major agents and smooth, continuous economic environments for the EU leasing market (leasing being one of many ways in which businesses finance their capital investments). Each national leasing market can be viewed as an agent, with a range of particular internal dynamics that gives it specific character (e.g. preference of national businesses in the use of leasing over time, expectation for future economic growth, attitudes towards other forms of financing investments etc). At the same time, a number of exogenous 'forces' also have an effect over each agent: forces such as the evolution of other national economies in close proximity, cross-border economic activity, pan-European taxation/regulation changes etc. By studying the leasing penetration in each national market (defined as the ratio of new yearly leasing volumes by the total yearly fixed capital formation in each economy) and comparing them with a measure of each economy's overall wealth (e.g. GDP per capita), Europe's national leasing markets fall into three clusters of agents: the first includes economies that are both large and wealthy (viewed in GDP/Capita terms) with a mature leasing market reaching high penetration levels. The second cluster includes economies that are wealthy and mature, but show very low leasing penetration levels. A third distinct cluster includes broadly the new EU entrants, i.e the smaller but high growth economies of Central and Eastern Europe, characterised by low GDP/Capita levels and at the same time exhibiting high leasing penetration levels. An Agent-Field model can be developed to map the dynamics that drive each cluster of economies, so as to help predict the direction that the third cluster of Europe's high growth economies can be expected to take, as its economies move towards higher prosperity levels. Within the scope of the work, it has been shown that the Agent-Field approach appears to be an intuitive rather than an abstract process in modeling economic systems. This intuitive process needs more understanding of the interactions between the economic environment and the agents within it as these elements represent the logic underlying the problem at hand rather than mathematical notation. The Agent-Field approach seems ontologically well founded for the growing field of agent-based computational economics.
Posts
