Dynamic models of multicriteria to combine variables: the use of parametric model and genetic algorithm in the study of the World Heritage area recognized by UNESCO in Pampulha, Brazil

Abstract

The present paper surveys dynamic models of multicriteria to combine variables using parametric model and genetic algorithm as a method of changing the adequacy level of variables in a multicriteria analysis (MCA). The aim is to simulate if-then scenarios of territorial occupation of commerce, housing and green areas. The case study is a MCA for the buffer zone of the modern assembly of Niemeyer in Pampulha region, Belo Horizonte, Brazil, declared World Heritage by UNESCO. The parametric model was developed in Grasshopper software. The level of adequacy/score of the territorial units to characterize attractiveness and vulnerabilities to land use change was defined by "knowledge-driven" in the layers: Safety Risks, Fragility in Infrastructure, Bus Stop and Centralities due to Interaction Potential. The land use change simulation "if-then" was defined by "objective-driven", due the use of fitness-function in genetic algorithm, with the goal to achieve the best distribution of land use changes, in order to result in a more balanced use of the territory (commerce, housing or vegetation), but also considering attractiveness and vulnerabilities defined by the characteristics of the neighborhoods (centralities, transportation, safety and fragilities in infrastructure). The parametric model generates “if-then” simulation, calculating an index of suitability for each territorial unit and changing the land use according to the objective-driven to be achieved in fitness-function.