Land use optimization allocation based on multi-agent genetic algorithm
Yuan Man,Liu Yaolin
DOI: https://doi.org/10.3969/j.issn.1002-6819.2014.01.025
2014-01-01
Abstract:Rapid industrialization and urbanization have resulted in massive land-use changes in China over the past few decades, and the optimization of land-use allocation is an important method to ensure national socioeconomic security and to achieve sustainable development. Optimization of land-use allocation is a complex spatial optimization problem that involves allocating various land uses to specific units within a region. A purely local simulation model or a global optimization model is insufficient to optimize regional land-use allocation, and the dichotomy between these two models has polarized our abilities to avoid their shortcomings and generate desirable land-use patterns. It is essential to bridge the gap between the two models through the combination of top-down and bottom-up approaches. This research proposes a coupled model for optimal land-use allocation by combining the modeling framework of a local simulation model with the computing framework of a global optimization model. This is achieved by integrating a MAS that simulates the behaviors of land-use stakeholders with regard to their choice of specific locations, with a GA that simultaneously evaluates and optimizes land-use patterns to meet various regional development objectives. The land-use spatial pattern, together with the socioeconomic properties in the GIS provided geographical, environmental, and other valuable information for use by the various land-use decision-making agents in the MAS. Due to its ability to simulate individual decision-making and the resulting interactions, the MAS has the potential to represent the aggregated outcomes of individual land-use allocation decisions. According to the hierarchy of regional land-use allocation in China, there are three major types of land-use stakeholder agents:the government, government departments, and residents. In accordance with regional objectives, the GA operates by algorithmically optimizing the land-use patterns that are grounded in the realm of MAS-simulated stakeholders’ decisions. The gap between the MAS and GA is bridged through developing an integrated agent-genetic evolutionary process, which combines the genetic operations of the GA with the decision-making behaviors of the MAS. The Caidian District of Wuhan, China, is an area of rapid economic growth and a sensitive ecological environment, making it our choice of study area to test the model. The model was expected to reasonably allocate land use across spatial units in accordance with multiple objectives and constraints, and to optimize land use in terms of the quantity structure and spatial pattern. We selected Pareto fronts of the solutions to analyze the optimal land-use spatial patterns. The results showed that the optimal pattern developed by the model improved on the economic output, spatial compactness, and carbon storage of the current pattern and promoted sustainable regional land-use development from the local scale to the regional scale. The land-use conversion improved the quality of the arable land, increased the forest coverage, and improved the environment in rural residential areas. The spatial allocation of developed urban and rural land meets the requirement of intensive and economical land use. The spatial constraints of the natural habitat reserves improve the natural environment and protect natural resources from human disturbance. In addition, the model provides a spatially explicit tool for generating alternative land-use spatial patterns in accordance with regional development strategies.