Land use optimization with genetic algorithm, using the methodology of evaluating ecosystem services
Evolutionary algorithms provide solutions for complex decision making problems with multiple objectives, which make them widely used tools in several fields. These methods determine near-optimal solutions in the case of opposing and sometimes non-linear objectives, for which land use allocation problems are good examples. This means that individuals occurring in the evolutionary process are grids of land use patterns, and steps in cyclical evolutionary processes (mutation, selection) result in more and more optimal land use allocation, which can be characterized e.g. by the weighted sum of pre-defined objective functions. In our contribution, we used the Boundary-based Fast Genetic Algorithm. The landscape ecological framework of the method is the evaluation of ecosystem services. The aim of the algorithm is the optimization on the land use type level, therefore, the evaluation is carried out partly with land use based proxy values, which is the simplest approach in ecosystem service evaluations. In our presentation, we show a case study from the Danube-Tisza Interfluve (Hungary). The land use conflict is characterized by the tradeoffs between the interests of agriculture (provisioning services) and the need for extensive land use forms (regulating and supporting services). Food provision and timber production were evaluated using agricultural statistics and wood production economic models, while the values of carbon sequestration were based on a targeted metaanalysis. The objective functions of landscape heterogeneity and ecological value are using landscape metrics as indicators. The model results highlight the need for the higher proportion of extensive land use forms in the study area.