Abstract: In the process of economic development, the unreasonable land allocation had led to deterioration of urban ecological environment which includes soil and water loss, farmland degradation and traffic jam. Therefore, optimal land use allocation has been regarded as one of the keys to the sustainable development of cities. Since some significant contradictions often exist between land development and nature conservation, land use allocation confronts great challenges to balance land use suitability, compactness and ecosystem services values.
In order to solve this problem, this paper aims to propose an ideal solution by using the multiobjective particle swarm optimization algorithm based on principle of pareto optimization. Six major land use types are arranged in study area which included cultivated land, forest land, wetland, waterbody, habitat and industry land. Three objective functions are defined: maximizing suitability, maximizing ecosystem service value, maximizing land use compactness. As the initial swarm was created, the values of the objective functions were calculated, and nondominated answers were preserved in an external archive. To prevent particle swarm sticking into local convergence, mutation operators were added to the particle swarms to promote the uniform distribution of particles. Combining with the constraints of the land policy, the Paretofront that meets the realistic requirements was finally formed in the threedimensional space. We selected Wujin district of Changzhou as study area, which had the land use maps, urban planning and socioeconomic data for the experiment of optimal land use allocation.
The results show that a uniform set of ParetoFrontiers can be obtained by applying the model, which means that each result is an optimal land use allocation. The optimized allocation improves the land suitability, ecosystem service value and land compactness by 9.96%, 4.27%, and 5.23% respectively. Compared with singleobjective model that can only produce one result, the model can obtain more optimized allocations with the same optimal level, different quantity structure and spatial characteristics, which can be better support for land use planning.