不规则邻域元胞自动机通过定义一定的邻域规则,将对中心元胞影响较大的邻域元胞进行识别与计算从而确定邻域形态与影响范围,与传统元胞自动机模型相同尺寸邻域形态相比,模拟更加真实有效。基于不规则邻域识别算法对元胞邻域范围进行划分,再通过遗传算法和多准则评价相结合获取转化规则参数,继而对大连市金石滩国家旅游度假区2004年和2010年土地利用变化进行模拟研究,通过比对分析以及Kappa系数检验法对模拟精度做一检验,研究模拟结果总体Kappa系数为81.62%,具有一定的可靠性,研究显示该模型在多地类碎小斑块之间的转化模拟具有一定的优势,对于模拟土地利用/覆盖变化模型具有一定的改进。 By defining a certain neighborhood rule, the cellular automaton with irregular neighborhoods can identify and calculate the neighboring cells that have a great influence on the central cell to determine the shape and influence range of the neighborhood, which is the same as the traditional cellular automaton model Compared to the size of the neighborhood morphology, the simulation is more realistic and effective. Based on the irregular neighborhood identification algorithm to divide the cell neighborhood, and then obtain the conversion rule parameters through the combination of genetic algorithm and multi-criteria evaluation. Then, the land use change of Jinshitan National Tourism Resort in Dalian in 2004 and 2010 Simulation study shows that the Kappa coefficient of the simulation results is 81.62%, which has certain reliability through comparison analysis and Kappa coefficient test. The research shows that the model has a good agreement among many small patches The conversion simulation has some advantages and has some improvements for simulating the land use / cover change model.