Resumen de Geoworldsim: a time-asynchronous, distributed and intelligent environment based geosimulation platform
Ander Pijoan Lamas
Imagine that the accessibility of the population to public infrastructures needs to be evaluated. A possible solution would be to calculate every distance and analyse which percentage of the population is in less than 300 meters. However, this solution does not take into account issues such as: the age distribution, functional diversity or people’s preferences when transiting the city. It is in these cases when it is necessary to go a step further and integrate Geographic Information Systems with other perspectives such as Multi-Agent Systems which represent the particular characteristics of each individual and their decision making processes. This integration is known as Geosimulation and builds more accurate simulations to model the physical reality together with social, demo graphic and economic components.
Geosimulations aim at modelling systems at the scale of individuals and entity-level units of the built environment and provides a way to simulate big amounts of agents interacting in a virtual geographic environment and endowed with spatial cognitive capabilities (perception, navigation, reasoning). This dissertation presents a new Geosimulation platform design and implementation that allows analysing and simulating different urban infrastructures. The platform manages to put into practice the latest theories in Multi-Agent Systems along with the new techniques in cloud computing and asynchronism. The proposed design is evaluated for three case studies; ubiquitous IoT, sustainable transport policies and resilience of the power grid. The methodologies presented, provide progress to their respective research areas by improving state-of-the-art techniques or designing new mechanisms.
Furthermore, by connecting these Geosimulations to the real world by sensors and actuators, the concept of mixed reality arises; simulations where changes in the real world are transferred to the virtual world through sensors and agents can influence the real world through actuators. Mixed realities allow developing distributed control systems, which not only take into account the physical reality and social preferences but also the state, where to deploy intelligent agents that provide services to citizens.
