Land use pattern optimization based on shuffled frog leaping algorithm

Abstract: Rational land use pattern has gained more and more attention, because it will optimize the allocation of regional resources and promote the regional flow of material and energy. However, most previous studies tend to focus on the optimization for number and structure, and ignore rational land use pattern optimization. How to unify number and structure and match target with the specific grid unit has become burning issues. In addition, there will be a large number of grid cells and the complicated calculation in the land space optimization, but the traditional methods do not meet the needs. The development of spatial information science provides an important support for the spatial decision making. With the development of the intelligent optimization algorithm combined with the GIS (geographic information system) technology for the modeling of the intelligent land use optimization, a number of land optimization models with high efficiency and strong solving appeared successively. Yet, most of these models cannot show microscopic changes of inland use decision-making process and reflect the evolution of microscopic space. So it is difficult to get a more precise allocation scheme of land resource. Traditional land optimization models, such as linear programming, multi-objective optimization, grey system and landscape ecology, cannot realize quantitative and spatial optimization simultaneously. In this paper, a land use spatial optimization model based on the shuffled frog leaping algorithm (SFLA) is proposed, which reduces the size of land grid cells in the intelligent optimization algorithm to the 30 m×30 m to get more accurate solution. The SFLA model, as a new intelligent optimization algorithm, has high computing performance and excellent global search capability, which can simulate the behavior of the frog population transfer within the group when searching for food. In the process of local searching, the frogs' individual information gets transmitted in group. After frog's merging, sorting, and regrouping in the various groups, the local information gets global exchange within the frog population. Local search and global information exchange continue to alternate until the convergence condition is satisfied. The spatial distribution of land pattern is simulated by the spatial distribution of frog swarm, and the process of searching for the optimal allocation of land is simulated by searching for the best frog in this model. Making the geographic grid unit of 30 m×30 m as the basic operating object, the output of the model corresponds to a feasible scheme of land use optimization. Taking the land use pattern of Lanzhou in 2014 as an example, choosing the land ecosystem's service value and the structure compactness as the optimization goal, the SFLA model is used to optimize the spatial pattern of land use in Lanzhou City in this paper. When the frogs jump to search for the optimal value, they try to improve the ecosystem as well as make land pattern concentrated, which effectively overcome the disadvantages of pursuing the quantity optimization and ignoring the rationality of the space. The ecosystem service value provided by various land types was 5.701×109 Yuan in 2014, and the optimized ecosystem service value is 5.802×109 Yuan. The land pattern standard compact degree was 0.37, and the optimized value can reach 0.47. After optimization, the area of grassland increases by 8 431 hm2, which increases ecosystem service value by 1.71×108 Yuan. The forest area increases by 1.453×105 hm2, which increases ecosystem value by 2.49×108 Yuan. The experimental results show that the SFLA model can simulate the spatial distribution of the land use and find out the optimal solution of the problem under the multi-objective control. Moreover, the model can unify number and structure of land use effectively, which has faster convergence speed and stronger search ability. The establishment of the model will provide a new way of intelligent optimization algorithm, which can be applied to solve the problem of land use optimization in smaller scale.