Using genetic algorithms to optimise land use suitability

Land-use planning is defined as the most appropriate utilization that would achieve the paramount benefit of protecting the resources. In this study, under environmentfriendliness objective, based on multi-agent genetic algorithms, was developed a geospatial model for the land use allocation. The model applied to solve the practical multi-objective spatial optimization allocation problems of land use in the core region of Menderjan Basin in Iran. The first task was studying the dominant of crops and economic suitability evaluation of land with the land evaluation framework developed by FAO, (1976-2007) using GIS. Second task is to determine the fitness function for the genetic algorithms. The third objective is to optimize the land use map using economic benefits. In the socioeconomic assessment of the Menderjan watershed; consultation with experts and the interview with local residents implemented. Different scenarios then arranged according to the land suitability classes. The Erosion Potential Method (EPM) used in erosion estimation and sediment yield of the study. The highest annual erosion rate belongs to the potato agricultural land use. Third scenario suggested in comparison to the economic views.In this research, based on both irrigation managements of the crops and water demands' model of crops would be developed and calculated which they integrated in RS and GIS environment. In the GAs Model, parent selected among the initial population. In fact, the initial population includes the land suitability analysis, land use/ land cover, which is extracted from RS and scenarios of land evaluation and crop suitability. To sum up, coding is remarkably based on objective function, which it has been great in cost/ benefit from all cultivating activities and obtained costs of land erosion. After calculating the fitness function, which it includes, cost and benefit matrix, cost of changing land uses together, offspring (the next generation) which are importantly generated. Selecting the offspring during the research has been based on their capability of elitism. This selection implemented according to the percentage of progressing, comparison and replacing in GAs programming. Finally, the land use and defined scenarios obtained as optimized output, which is a dynamic model in this study. The results shows; the major limitations regarding to wheat in this region is related to the topography. 28.6% of the land has severe topographic limitations. The most suitable class is S2 for Potato. The limitation of this suitability class majorly is soil properties. Results of Almond land suitability analysis shows, the most extensive land is in the moderate limitation class. The main limitation is properties of the soil and climate. After doing the related analyses, it has been achieved that the water consumption (water demand) for wheat in May had the most consumption of water and April and June comes afterward. Potato in July has more water consumption and after that August, September, June and May. The erosion potential categories determined that heavy and severe class covered 35% of the area. Land use/ Land Cover is obtained by satellite image processing that the overall kappa of the classification is 87.4% and the overall accuracy is 89.6%. As it has mapped, the Irrigated area is 4689 ha. According to the results of the GAs Programme and the produced graphs in evaluating the best solutions, it has been recognized that after 25 frequencies there is not any intensify change, which it happened in the optimized beneficial value, so, extra reiteration has not influence in the possible better answer. The final optimized benefit is 12*1011.

Preview Text ITMA 2012 14 IR.pdf Download (2MB) | Preview

Actions (login required)

View Item