Divisão de Observação da Terra e Geoinformática

Essa página está em modo somente de leitura. Você pode visualizar a fonte, mas não alterá-la. Informe-se com o administrador do Wiki, caso você ache que isso está incorreto.
====== An Architecture for Agent-based Modelling and Simulation of Geospatial Phenomena ======

|P. R. Andrade, A. M. V. Monteiro, G. Camara, T. G. S. Carneiro|

[[http://cress.soc.surrey.ac.uk/essa2009/|6th European Social Simulation Association Conference]]


====review 1====

TITLE: An Architecture for Agent-based Modelling and Simulation of Geospatial Phenomena \\
OVERALL RATING: -2 (reject)  \\
Importance of work/results: 2 (below average) \\
 
REVIEW: This is a technical note rather than a conference paper. It describes,
as the title suggests, a toolkit for incorporating GIS data into
agent-based models. As such it is of some interest, but no specific
model is described, and no social phenomena are discussed. As such, I
am not convinced it is suitable for this conference. Rather, it should
be submitted to a conference or workshop focused on methodological
issues, or on the interface between agent-based mondels and geospatial
data. If the paper started from a specific modeling problem or
problems and described how the approach described contributed to their
solution it would be acceptable, but although the concluding remarks
section refers to two such specific problems (incidentally removing the
anonymity of the paper), they are not described. 
For the most part the paper is clear, but depends too much on the
reader having read, or being willing to read, other work in order to
understand it. 


====review 2====

PAPER: 48 \\
TITLE: An Architecture for Agent-based Modelling and Simulation of Geospatial Phenomena \\
OVERALL RATING: 0 (borderline paper)  \\
Importance of work/results: 2 (below average) \\
 
REVIREW:The problem with this paper is that the "challenges" or "difficulties" it tries to address are not clearly stated. As I 
understand it, it proposes a  way to represent agents and their relations with geospatial data, rather than "simply" 
using geospatial data as a "landscape" for ABM models. But why should agent-based models that use geospatial data 
be represented using geospatial data ? What is the problem in considering the environment and the way it is encoded 
as a "landscape" ? Why should we express everything in terms of geospatial data ? These questions are quite important 
because the "environment" (as well as the agents) can be described using many different techniques and that this 
heterogeneity should be, in my opinion, preserved even if they create computational difficulties. 


====review 3====

PAPER: 48 \\
TITLE: An Architecture for Agent-based Modelling and Simulation of Geospatial Phenomena \\
OVERALL RATING: 2 (accept)   \\
Importance of work/results: 3 (average) \\
 
REVIEW: The paper presents an interesting approach for the integration of GIS and agent based toolkits. The approach is clearly presented but in order to effectively conclude the paper, an application example must be presented, especially if, as the authors said in the concluding paragraph, they have the possibility to make comparisons with other models of the same target made following other approaches.