|C.R. García-Alonso, L.M. Pérez-Naranjo
|multi-objective evolutionary algorithm, hierarchical Bayesian model, spatial analysis, financial risk in horticultural farms
Horticulture is the most important sub-sector in Andalusian agriculture.
The geographical identification of financially compromised and/or sustainable areas can be strategic for determining spatially structural and political measures.
Different methodologies have been used to detect highly (positively or negatively) auto-correlated zones in space (hot-spots) but all of them find it quite difficult to precisely pinpoint the spatial units included.
This paper describes a Multi-Objective Evolutionary Algorithm (MOEA) designed to identify hot-spots at municipal level based on the Bayesian Conditional Auto-regressive (CAR) model.
Our MOEA (SPEA2 model) evaluates the probability each spatial unit has of belonging to a potential hot-spot and results can be represented on a map.
Hot-spots were identified by optimizing the spatial distribution of Bayesian risks, minimizing their standard deviations and minimizing the minimum path (distances) that links all municipality capitals included in the potential hot-spot.
The results lead to a better understanding of problems related to rural sustainability.
Download Adobe Acrobat Reader (free software to read PDF files)
Hosted by KU Leuven LIBIS