Spatial Optimal Allocation Of Land Use Based On Multi-Objective Genetic Algorithm In Beijing-Tianjin-Hebei

It is necessary to consider various factors such as economy,society and ecology in the optimization of land use.Therefore,land use optimization is a typical multi-objective optimization problem.For the traditional genetic algorithm,the multi-objective optimization problem is transformed into a single-objective problem by the fitness function,the obtained solution set is relatively simple and easy to fall into the local optimal solution,the convergence speed is slow.Aiming at the disadvantages of the traditional genetic algorithm,this paper adopts NSGA2 algorithm with the elite strategy and the fast non-dominated sorting mechanism.Genetic algorithm(NSGA2algorithm)has the advantages of better solution set and lower complexity.This paper picked the Beijing-Tianjin-Hebei urban group as an example.Based on the current status of land use quantitative structure and spatial pattern in 2015,this paper selected three types objectives(economic,environmental and land),and used genetic algorithms,cellular automation model(CA),and geography information system(GIS),constructed the NSGA2-multi-objective model,finally,verified the model operation results.The main research contents and conclusions are as follows:(1)Land use structure and pattern of Beijing-Tianjin-Hebei.In terms of quantitative structure,the area of cultivated land in Beijing-Tianjin-Hebei accounted for about half of the total area(49.43%),the second is forest land(20.85%)and the third is grassland(16.49%);then,the construction land area grew fastest(7.7%)in the past 10 years,the area of cultivated land decreased the most(1023.0km~2).From the perspective of spatial pattern,the land use fragmentation degree of Beijing-Tianjin-Hebei was low(patch density<1),and the overall distribution was relatively uniform(SHEI=0.747),but construction land is the most dispersed land type(aggregation index=35.915),and the utilization is not economical and intensive enough.(2)Obtained the optimal solution set of the Beijing-Tianjin-Hebei land use plan based on the NSGA2-multi-objective model.In the optimization process,by 2020,the objective value of the one maximum indicator GDP decreased gradually,the objective value of the four minimum indicators of non-point source pollution(NPS),aerosol optical depth(AOD),the cost of changing from the status quo and the incompatibility between land uses was gradually decreased too,but the predicted value of the maximum indicator GDP was still higher than the GDP value obtained by the unit land area GDP forecasting method;all the plans met the minimum area constraint of construction land,cultivated land and waters;In the Pareto optimal solution set(10 plans)derived from the NSGA2-multi-objective model,the construction land area had a certain amount of increase(345-1231km~2),and the cultivated land and grassland area had a certain amount of reduction(cultivated land970-1541km~2;Grassland 48-94km~2),the water area was changeless or had a small increase(0-269km~2),and the forest land and unused land of the seven plans had increased(forest land 108-474km~2;unused land 59-536km~2).(3)Verified the effectiveness of the NSGA2-multi-objective model to solve the land use optimization problem.Compared with the 2015 land use status,CA simulation plan and NSGA2-multi-objective model optimization plans,it was found that the Pareto optimal plans obtained by the NSGA2-multi-objective model had the higher GDP and the lower NPS,AOD and the incompatibility than the 2015 land use status and CA simulation plan,therefore,these indicators’objective values were superior to the 2015 land use status and CA simulation plan.Only the cost of changing from the status quo was higher than the CA simulation plan,and this indicators’objective values were inferior to the 2015land use status and CA simulation plan.From the landscape pattern index of the Pareto optimal plans,the patch density,aggregation index,SHDI and SHEI of all plans were superior to the land use status and CA simulation plan,the shape index of 8 plans was better than the land use status and CA simulation plan;the operation results illustrated the effectiveness of NSGA2-multi-objective model for land use spatial optimization.